EU strikes deal to regulate ChatGPT, AI tech | Hacker News Reader

EU strikes deal to regulate ChatGPT, AI tech

bloomberg.com

218 pointshelsinkiandrew464 comments

> By 2022, there were examples that AI was actively harming people. In a Dutch scandal, decisions made by algorithms were linked to families being forcibly separated from their children, while students studying remotely alleged that AI systems discriminated against them based on the color of their skin.

These were not AIs but rule-based systems with if elses. The current legislation is bad, because it tries to regulate the technology, not how it is used. Dutch government people should be responsible for the decisions they or their system produces, AI or no AI.

> The current legislation is bad, because it tries to regulate the technology, not how it is used.

That's not what the EU regulation does. As an example, High Risk Machine Translation will be regulated no matter how you do it. The standard underlying the EU's AI Act explicit mention that AI system can be anything from rule-based to machine learning and mention some other AI methods. The regulation is on task/usage not on the technology.

Edit: "High Risk" in my example refers to the usage

Thank you for the clarification!

> AI system can be anything from rule-based to machine learning

Then why call it AI?

Why call anything AI?

It's become commonly accepted to mean any algorithmic system. It's probably better to use an accepted umbrella term rather than multiple specific scientific ones in this case.

Fuzzy and misused words are pushed to muddy the semantics. Actual AI doesn't exist.

If/else, case, and while statements are basic building blocks of automation and computational logic. Intelligence is far from anything we've done with computers. It is a parlor trick looking for money right now.

> Actual AI doesn't exist.

The word "actual" is doing a lot of lifting there. Artificial intelligence has existed for decades. Presumably you mean "artificial general intelligence", whatever that's even supposed to mean (it's subjective, and for many people it's a matter of literal religious faith that humans have something called souls which machines can never possibly emulate, and therefore "actual AI" is literally impossible as far as they're concerned.) Well anyway... if you want to say AGI, just say AGI. AI is more general than that and the term is well established to encompass a great many methods and kinds of systems. Many kinds of AI are very primitive and seem trivial today; they are nonetheless "actual AI".

I will not bend to the sloppy use of words to misrepresent technology. My words are not your putty to mold.

I get what you are saying.

With a technical audience, it is better to use actual terminology.

However the general public does not have the knowledge to understand these. These are the people using ChatGPT and for them it is AI. We can correct them, but it's likely they won't care.

The general public is being misled with overwrought appeals to emotion and misused terms applied in a deliberately overbroad manner. AI may contain algorithms but an algorithm is not AI.

Social media feed algorithms are bad, and the public is awakening to the damage they cause. But they are not AI. As Yan LeCun pointed out recently, AI is actually the solution to the problems created by "algorithms."

It's not "sloppy use of words", both the term and the practice of AI have been around for far longer than today's probabilistic models. In particular, decision trees and similar are some of the simplest forms of AI.

The difference between an AI and a "regular" bunch of if/else statements is, in my opinion, the ease of adding new rules to the system. That's why something like Prolog counts as AI.

Intelligence is the capability to form (predictive) models (to solve problems through evaluating strategies through the aforementioned model). General intelligence is the ability to do it over an arbitrary range of problems.

Even an if-else tree is a model. Building the model based on data leads to high predictive power.

AI is an okay term for this IMHO.

"If-else rule-based systems" aka expert systems are an example of symbolic artificial intelligence e.g. "Good Old Fashioned Artificial Intelligence". Just because it doesn't use the newest most trendy methods doesn't mean it isn't AI, and even if you object to applying the AI label to these kind of systems, these systems are nevertheless valid targets for regulation.

"Good Old Fashioned Artificial Intelligence" - https://en.wikipedia.org/wiki/GOFAI

"In artificial intelligence, an expert system is a computer system emulating the decision-making ability of a human expert.[1] Expert systems are designed to solve complex problems by reasoning through bodies of knowledge, represented mainly as if–then rules" - https://en.wikipedia.org/wiki/Expert_system

"Symbolic AI used tools such as logic programming, production rules, semantic nets and frames, and it developed applications such as knowledge-based systems (in particular, expert systems), [...]" - https://en.wikipedia.org/wiki/Symbolic_artificial_intelligen...

Because "AI" is a well-established academic field, and the definition by your quoted comment is in line with this.

It's a very broad umbrella term, machine (and deep) learning are only subsets - although the main drivers these days.

Marketing.

An acronym with the word "statistics" would not have been able to sell the false sense of precision or superior deduction or the 'entity able to think' that is implied by putting "AI" in the sentence.

Such marketing is hijacking what the general public has come to know over the past two centuries as AI through sci-fi novels and films, which now they wants to be rename to Artificial General Intelligence, AGI ?

One would expect they edit all the sci-fi novels and movies around the history for to avoid the AI acronym confusion, but in such case the marketing would stop being effective, because the quid is within my first sentence, they sell the false sense of precision or superior deduction, when in fact there are only statistical models. Pure oxymoron, as statistics in itself sounds like an acronym for "lack of precision".

This is sort of my point - one of the options is "rule based system". That doesn't need to be statistical at all. It doesn't look like what people would think of as AI at all. I don't see why it isn't just "you can't use computers for this" and AI is an implementation detail.

> It doesn't look like what people would think of as AI at all.

If by “people” you mean specifically the subset of people who have technical knowledge or who read HN, then sure, but in the wider population, anything that seems intelligent that’s fine by a computer is AI, they don’t care if it’s using statistical techniques, a random number generator or a bunch of conditionals. Laypeople don’t know or care what happens under the hood.

I once heard a conversation between an old boss of mine and an investor where my boss was saying that the software did calculations X and Y automatically and the investor responded with “so it’s AI”. My thought was “wait what? No it’s not” but my boss said something like “AI is anything that people perceive as intelligence thats artificial”. Historically, exist systems were seen as AI and they weren’t necessarily statistics based. That’s why we have more technical terms like machine learning.

But I agree, that laws should be far more specific about what they mean and “you can use a computer for this” would be better.m, if that’s what they really mean.

By such logic a company could call wood steam engines neutron-free nuclear fission engines, as they convert the water vapour into mechanical energy. It would be to change and/or to blend the properties of any technical word or concept freely, and legitimately, for to commerce, marketing promotion or any purpose. If generalized I guess it would promote and increase the misleading advertising cases exponentially.

( If I can humbly comment about what happened in the conversation you heard, it seems that your boss eluded an "it is not Machine Learning, it is better" and blended an answer for to avoid confronting or disappointing the investor some way )

I spoke to him about it after and he said something like: if it looks intelligent to the user, and it’s done by a computer, the user sees it as artificial intelligence, since they don’t care about the underlying tech that drives it. Just how startups often manually process the first few customers requests or get mechanical Turk to do something. Users just see the end result and don’t really know or care if it was implemented by a mess of if statements, machine learning, or magical fairy dust. They just see a machine acting in ways that seem intelligent.

Of course that very much muddies and dilutes the terminology, but AI was always an imprecise and badly defined term and, as was pointed out, earlier AI techniques wouldn’t be seen as that nowadays: by todays standard, I wouldn’t consider expert systems or minmax tree search or whatever as AI, but once upon a time, when the term was coined, it certainly was.

Because that’s what it has been called since the term was coined in the 1950s.

I think its meaning may be shifting, but https://en.wikipedia.org/wiki/Machine_learning#Artificial_in... still shows it as a broader term than “machine learning”, including expert systems.

Also note that decision tree learning (https://en.wikipedia.org/wiki/Decision_tree_learning) basically is a search for a good rule-based system.

Because an expert system is an AI

Because it's still AI. If I write a simple program "if black, no scholarship" it should still face the same regulations as the latest LLM evaluating scholarships.

What we call "AI" today we called algorithms yesteryear and applications yesterdecade.

It's all marketing, I think EU is right to accurately use the "AI" marketing term for what it is.

> It's all marketing, I think EU is right to accurately use the "AI" marketing term for what it is.

I think EU calling it "AI" is fine today. I'm more worried about the future. In 5 years when the hype word for this class of tools becomes Tchounka. The world will say that EU is late and they need to pass a Tchounka law. While the reality is that EU passed an automated data processing law and doesn't care about how it's done.

> In 5 years when the hype word for this class of tools becomes Tchounka.

You could read Annex I of the law which details what EU calls AI instead of speculating.

And yes, if some new way comes along the law can, and will be, amended

AI has always been used to describe plain algorithmical things as well. See video games for example.

But ultimately everything in life can be thought of as algorithm so it really doesn't matter. If it's artificial and it can solve problems beyond simple computations, it is AI.

Humans are algorithms as well, just very complex and nuanced ones that we have a lot of trouble reverse engineering.

I would say that a complex hardcoded algorithm that is used for a racing game is a stronger AI than a simple, small neural network for one specific narrow task.

Intelligence = ability to solve problems, AI = something that is artificial and can solve problems beyond being a simple tool.

Sounds what something someone might want to call an expert system?

> In artificial intelligence, an expert system is a computer system emulating the decision-making ability of a human expert.[1] Expert systems are designed to solve complex problems by reasoning through bodies of knowledge, represented mainly as if–then rules rather than through conventional procedural code

Do you have any reference for this claim, or are you guessing? It was reported that the algorithm was a Gradient Boosting Machine by investigators who gained access to the code.

https://www.lighthousereports.com/suspicion-machines-methodo...

Well if you torture the analogy enough, you could argue a tree based model is a lot of if/else statements!

According to the EU proposal, this is what they mean by AI:

(a)Machine learning approaches, including supervised, unsupervised and reinforcement learning, using a wide variety of methods including deep learning;

(b)Logic- and knowledge-based approaches, including knowledge representation, inductive (logic) programming, knowledge bases, inference and deductive engines, (symbolic) reasoning and expert systems;

(c)Statistical approaches, Bayesian estimation, search and optimization methods.

So rule-based systems (b) are considered AI for the sake of regulation.

Aren't all current machine learning approaches statistical approaches using Bayesian estimation?

What do you mean by “current”?

Not really.. they're often statistically based, but don't need to use Bayes' Theorem or be based on a Bayes Network. They are driven by expectation maximization and stochastic gradient descent, mainly. An autodifferentiation back-propagates the parameter update through a neural network model, but the network doesn't necessarily represent a Bayes Network.

If that were the case, that wouldn't be just one of the covered concepts in the proposed law. We should not assume world-shaking incompetence.

So every program ever made?

Yes.

That's the idea, yes.

It is not really about the techniques used but about what you are using these programs for. If law enforcement used, say, the average brightness of a picture of someone's face as a basis for suspicion, it could hardly be called AI, but it is nevertheless a highly concerning usage of technology and shouldn't be given a pass just because it is not advanced enough.

I'm sorry but I cannot see how to view this charitably. To me it clearly seems to be a way to enact further unjustified regulation on something that would otherwise be hard to create public support for, by using big new scary words like AI, even though it has massive overreach into more traditional programs.

If you tell people the computer program that profiled them as a probable scumbag and denied them credit or a job was using methods from the 1980s instead of the newest neural nets, that won't change how they feel about it. Objecting to regulation of such systems because you think the term "AI" should be reserved for the newest and trendiest systems very "HN" and misses the forest for the trees. The EU is rightly concerned about the ways that computers empowered to make important decisions might impact people's lives might perpetrate injustice; that's what is important here. Not some whiny semantic nitpicks about what to call those programs.

If you tell people it was a program using methods from the 1980s instead of the newest neural nets, it likely would change how they feel about the newest neural nets. Clarity is the only way to avoid cutting off our nose to spite our face.

But don't you think that there should be regulation saying that delegating decisions to a technology system should make you fully responsible, and that growing capability of these systems make it more important to clarify that any "mistrusted the machine by accident" argument is null and void?

For me, it's not important if the legal definition requires this term. It's not like automation and computer programs are being banned.

It's the decision part that justifies this legal definition IMO.

It's like the difference between a human or an automated car running over people. Quite literally even.

> But don't you think that there should be regulation saying that delegating decisions to a technology system

What is a "decision"? My computer is making billions of decisions per second.

If nothing else, I could see this crippling FOSS. No one is going to want to be "responsible" to some EU bureaucrat pursuing a personal political agenda, certainly not for work you've done for free.

Yeah I wanted to expand but got tired:

Of course any computation can be considered a decision in some way. But this is conflating a narrow and IT-specific meaning of that word with legalese and philosophy, or my intended meaning of the word.

This reminds me of the ancient catastrophic Therac-25 software bug, and other such cases.

Maybe this case is a good example for thinking about what part of the responsibility is on the side of the operator, apart from the obvious failure of the implementer.

For more modern examples involving actual ML, look at Meta, TikTok and their recommendations: where is the line to draw, what excuses are allowed, when the outcome is obviously negative, and the algorithm claims to fulfill a goal?

It doesn't matter if it's a rule-based system without "intelligence" or ML.

What matters is the responsibility of the human operator.

And making assumptions about correctness.

Humans make egregious errors as well, but the kind of errors AI causes are a significant concern where current legislation is insufficient.

It's one thing to have a bug in your airplane controller code or whatever.

It's another to knowingly accept malicious errors, either unpredictable or even intentional, but without proper responsibility?

> It's another to knowingly accept malicious errors, either unpredictable or even intentional, but without proper responsibility?

Is this law making the users of software responsible (e.g. law enforcement, government departments) ? It seems to me to make authors of software responsible, absolving the users. Not the other way around.

They keep the law secret, of course, so anyone can claim what they want, but the article talks directly about the accountability of OpenAI. It seems to focus on rules on authors of AI software, presumably to mostly absolve governments and users of that software. "Rules around generative AI", "transparency requirements for any developer of a large language model" ... nothing that would make governments responsible for abusing AI software. Which is strange, because that's the concern the last paragraph of the article focuses on.

I must say ... this "complete ban" voted by the EU parliament last spring, doesn't seem to have stopped governments from using live facial recognition [1].

[1] https://www.euronews.com/next/2023/02/21/new-french-facial-r...

> It seems to me to make authors of software responsible, absolving the users.

Only the authors of foundational models.

> They keep the law secret, of course

You mean this secret law: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52... that people have actually read https://softwarecrisis.dev/letters/the-truth-about-the-eu-ac... ?

> Developers (not deployers) of foundation models need to register their models, with documentation, prior to making it available on the market or as a service.

No, thanks. Utterly insane, we are sliding backwards into the "code is a munitions export" territory that I so hoped we'd escaped after the 90s.

> If you take a foundation model, fine-tune it for a specialised purpose, and deploy it as a part of your software, it won’t count as a foundation model, and you’ll probably be fine, as long as the original provider of the foundation model was compliant.

This interpretation would render the whole exercise pointless, there is quite a blurry line between "fine-tuning" and "training".

> Utterly insane, we are sliding backwards into the "code is a munitions export"

Of course not. This goes hand in hand with such things as "being accountable". You can't spit out a black box that no one knows what it does and how it was trained? Too bad for you.

Perhaps this will lead to fewer things like this: https://news.ycombinator.com/item?id=38595751

> You can't spit out a black box that no one knows what it does and how it was trained?

What if I want to? Who on earth has the ethical authority to claim the right to control my ability to release a set of weights? Only what is done with those weights should be legislated, and extremely conservatively.

Maybe a lot of software developers got into this career for the money, or because they like solving problems. But for me, politics inseparable from software engineering. Politics are why I devoted myself to the craft. This is the exact kind of situation in which my knowledge and skill become tools of protest.

I don't think this law is meant to stop these things at all. It's meant to make sure that companies and governments who do this and use models as an excuse go free and have someone to blame if they get caught.

> Who on earth has the ethical authority to claim the right to control my ability

Do you also complain about so many other things that "unethically curb your abilities"?

> Only what is done with those weights should be legislated, and extremely conservatively.

Ah yes, let's regulate this black box with no insight into what it does, and only guess at its possible outcome. Whatever can go wrong?

Oh, we know what can go wrong, because we've had multiple issues with algorithms going wrong.

> Do you also complain about so many other things that "unethically curb your abilities"?

You curiously omitted the last part of the sentence: "...to release a set of weights". Please don't pretend that I was speaking about anything else, and don't overgeneralize my statements; that becomes a straw man argument. Believe it or not, some laws are unethical. I am happy to provide examples.

It's a case-by-case basis which involves evaluating the overall impact on human rights for all parties involved.

The ideal scenario is one where all rights are preserved under good faith, and publishing/owning models is treated no different than any other software project, while the actual use of such software continues to be subject to existing laws. In this case, can strengthen consumer rights without weakening developer rights.

> Ah yes, let's regulate this black box with no insight into what it does, and only guess at its possible outcome. Whatever can go wrong?

I specifically said we should not be legislating weights, so I'm confused about which point you are trying to make. Weights are the black box. Company policy, employee behavior, and business logic are not, and are accessible for scrutiny by the courts if needed. So no, let's not regulate the existence of software, which sets an incredibly dark precedent for digital sovereignty.

> while the actual use of such software continues to be subject to existing laws. In this case, can strengthen consumer rights without weakening developer rights.

Emphasis mine

--- start quote ---

ANNEX IV

TECHNICAL DOCUMENTATION referred to in Article 11(1)

...

2. A detailed description of the elements of the AI system and of the process for its development, including:

...

where relevant, the data requirements in terms of datasheets describing the training methodologies and techniques and the training data sets used, including information about the provenance of those data sets, their scope and main characteristics; how the data was obtained and selected; labelling procedures (e.g. for supervised learning), data cleaning methodologies (e.g. outliers detection);

--- end quote ---

So let's see Article 11(1)

--- start quote ---

The technical documentation of a high-risk AI system shall be drawn up before that system is placed on the market or put into service and shall be kept up-to date.

--- end quote ---

Conclusion: all you're doing is spreading FUD.

I've made no claims as to the nature of the law, I've only asked questions about particulars and responded to the answers. If anyone is spreading FUD, it isn't me.

> > Developers (not deployers) of foundation models need to register their models, with documentation, prior to making it available on the market or as a service.

> No, thanks. Utterly insane, we are sliding backwards into the "code is a munitions export" territory that I so hoped we'd escaped after the 90s.

I'm not sure if that's what the legislation is about though? I hoped for it to be more about legal responsibility for the results of operating AI services, considering they pose difficult legal challenges, at least I suppose so, regarding responsibility.

I admit that I have not studied the law in detail, and in case it wasn't clear, I am Not A Lawyer.

If the law instead is about regulating the release of model weights, I will understand this and it would of course disappoint me.

> If the law instead is about regulating the release of model weights, I will understand this and it would of course disappoint me.

Of course it's not about that. It's pure unadulterated FUD. The law is about many things:

https://softwarecrisis.dev/letters/the-truth-about-the-eu-ac...

Nobody cares about your CMS, online shop, code collaboration tool, etc.

If your software however makes life-altering decisions for its users - such as targeting people for investigation, deciding on asylum requests, etc. - then you have to be responsible for the decisions your software makes. You can't hide behind "I don't know how the AI works, but it decided that person X is likely a criminal" - nope, you need to be able to explain the reasoning process behind this, because the system is your responsibility.

> If nothing else, I could see this crippling FOSS.

If anything, this should strengthen FOSS, because one of stipulations is "document your foundational models and training sets"

Creating regulations for hypothetical scenarios and possibilities you can invent in your head is silly.

Give me real examples of IRL harm. And most importantly give me real examples of how exactly state intervention directly solves those problems.

Otherwise this is just a philosophical debate mixed with prepper type fear of what could happen.

I assume one of the core idea would be to protect individual made by “blackbox” (complex or non complex) algorithm (e.g. used by banks and other companies) without any direct human involvement.

Basically banning companies from just saying “ Computer says no”..

The anxiety for regulating absolutely everything in EU is starting to be harmful.

We see mountains and mountains of regulations, some of ehich are really harmful (especially in the primary sector) and that I think are not well-intentioned.

There is a full agenda of control-everything that I do not find healthy for the average citizen.

> Give me real examples of IRL harm.

China's social scoring?

Racial profiling in government services? https://www.amnesty.org/en/latest/news/2021/10/xenophobic-ma...

Rejecting qualified applicants? https://www.eeoc.gov/newsroom/eeoc-sues-itutorgroup-age-disc...

Racial bias skipping patients in healthcare? https://www.scientificamerican.com/article/racial-bias-found...

Recruiting ignoring women? https://www.reuters.com/article/us-amazon-com-jobs-automatio...

Wrongly issued debts? https://en.wikipedia.org/wiki/Robodebt_scheme

and so on and so forth

> And most importantly give me real examples of how exactly state intervention directly solves those problems.

Like they solve about a billion others of problems daily, and you have no issues with government interventions.

> Otherwise this is just a philosophical debate mixed with prepper type fear of what could happen.

It's only philosophical if you willingly ignore the world around you

I gave an example right at the end of my comment (but might be I have edited it, so no fret intended). Self-driving cars.

Another commenter in this thread gave more examples, which further underline what I originally meant.

Physical harm means misclassified images/persons/posts/lifes/situations/... with the classification taken as gospel, in self-proclaimed good faith. Content moderation, credit scoring, police, the whole "new" generative space. — a lot of dangerous possibilities have opened.

All of them share the commonly accepted concept of "an a AI making a decision" (in common language).

This is another level of reliance, even if gradually, from computer systems and software in general.

I am not denying that complicated liability questions exist about that too.

> That's the idea, yes.

Hmmm... if that's the case, I expect to see a lot of software taken off the EU market, whether by simply removing it from the app stores for Euro Zone or other means.

This is gonna be the new GPDR.

> This is gonna be the new GPDR

We should certainly hope so given how much that forced a long overdue reckoning for the data collection and sales markets.

Well, no. Outfits like Facebook and Amazon can afford to hire attorneys to keep them in compliance with GPDR (or, more accurately, skirt around it).

Small website operators can't do that. Accordingly, many have banned EU IP addresses altogether.

The basic phenomenon is called "regulatory capture" -- the large player counterintuitively does not object to onerous regulations, because their smaller competitors can't afford to comply with them.

Happens all the time, in many different sectors of the economy.

With regard to FOSS, I certainly wouldn't work on anything for free that was likely to get me enmeshed in the legal system of another country.

If this legislation is as described above, someone's probably working on a FOSS license right now that allows the software to be used everywhere except the EU.

Complying with GDPR shouldn’t be that hard at all for most businesses if they introduce some basic procedures for handling/storing user data and most importantly don’t want to transfer/sell it to third parties

> can afford to hire attorneys to keep them in compliance with GPDR

Compliance with GDPR is trivial for the absolute vast majority of businesses. Here's how GitHub does it: https://github.blog/2020-12-17-no-cookie-for-you/

--- start quote ---

At GitHub, we want to protect developer privacy, and we find cookie banners quite irritating, so we decided to look for a solution. After a brief search, we found one: just don’t use any non-essential cookies. Pretty simple, really.

So, we have removed all non-essential cookies from GitHub, and visiting our website does not send any information to third-party analytics services. (And of course GitHub still does not use any cookies to display ads, or track you across other sites.)

--- end quote ---

> Compliance with GDPR is trivial for the absolute vast majority of businesses.

Nonsense.

> Here's how GitHub does it

That's nice. GitHub is a billion dollar Microsoft subsidiary with a gigantic legal team on speed dial.

What does "non-essential" mean here? Personally, I'm not willing to gamble that my interpretation of "non-essential" matches the interpretation of some random EU bureaucrat.

> Nonsense

It's absolute truth

> That's nice. GitHub is a billion dollar Microsoft subsidiary with a gigantic legal team on speed dial.

And instead of third-party tracking, selling your data to the highest bidder, and pestering you with miles-long cookie popups with hundreds of "partners", they don't do any of that.

So, how is this impossible for any other business that doesn't have those lawyers?

> What does "non-essential" mean here?

The law has been around for 7 years now. If you still ask this question and assume that you need to spend billions on lawyers to do the easy sensible thing, it only means that you don't care about answers.

> Well, no. Outfits like Facebook and Amazon can afford to hire attorneys to keep them in compliance with GPDR (or, more accurately, skirt around it). > Small website operators can't do that. Accordingly, many have banned EU IP addresses altogether.

It's much easier for small sites: not collecting data is free and if they aren't trying to resell it in ways which they don't want their customers to know about, it's easy enough to have a simple privacy policy. Those obtrusive banners are a political choice trying to make it look like GDPR compliance is onerous, but if you're not in the ad-tech business that's really not so hard since all of your cookies are essential.

GDPR did really good things for the market as a whole. The number of (inevitably US) websites that made a big stink either didn't take the time to understand what they needed to do (perhaps an hour for an average website by a competent engineer), or are so incredibly allergic to the implication that they must act ethically that they decided to say "fuck this". For these cases, the world is better off if they don't exist.

If you build a system and cannot live with the fact that a human must make the final estimation and be responsible for it's decisions, you should not be building a system that judges people.

> The number of (inevitably US) websites that made a big stink either didn't take the time to understand what they needed to do (perhaps an hour for an average website by a competent engineer), or are so incredibly allergic to the implication that they must act ethically

Or just decided that complying with EU security theatre administered by an unaccountable bureaucracy wasn't worth an hour of their time and just banned the Euros outright.

I agree with your point the users are at fault, not the technology itself. It's the govt which was separating families wrongly and the schools who discriminated who should suffer, not the companies making the tech.

That being said, this is a problem which regulation can be made to help solve. For example, a rule saying that any sold or licensed LLM to businesses needs to be run against a set of cleaning / filtering models (like the recently-released Purple Llama) before being packed up and sold / hosted. Maybe there can be an exception for individuals and exempt orgs (charities and religions may get an exemption, I could imagine).

> These were not AIs but rule-based systems with if elses.

What is fashionable to call AI these days are ultimately rule-based systems with if elses. It just so happens that it's a bazillion of those, and they are created by gradient descent.

And that they're not if-elses but nonlinear functions.

You mean functions such as:

   if x > 0:
       return x
   else:
       return 0

?

That's RELU, which clearly belongs to the realm of AI. Dangerous stuff.

It's almost like a chose the example on purpose ;)

That's one example of a nonlinear function, yes. You also have (eˣ)/(1+eˣ), or (e²ˣ-1)/(e²ˣ+1), or a host of others which don't contain anything if-like at all.

My somewhat snarky point was that there is no magic distinction such as:

> These were not AIs but rule-based systems with if elses.

I believe I already made my point with the ReLU example, if you don't want to concede it that is also ok.

I don’t know what point you’re trying to make. Neural nets are fundamentally different from rule-based systems. The fact that one of the non-linear functions which is sometimes used in neural nets can be implemented using an if statement seems utterly irrelevant.

The Alabama of Europe

As someone living in Europe, I don't know if I should laugh or cry.

The upside: From an investing perspective, the more regulations the EU comes up with, the more US software companies are a safe bet.

The downside: Europe missed the internet and now will miss AI. How will that affect quality of life here?

Is nobody in the EU bringing up the question why every single European citizen is dependent on US hardware, US software and US services? How we got here and where this will lead to, when software will not only be our "bicycle of the mind" but literally the car that drives us around, the nurse that cares for our health and the craftsman that builds our home?

why every single European citizen is dependent on US hardware, US software and US services

Mostly because US companies keep buying their EU competitors. Nokia, Skype, Spotify, ARM, ASML and countless others all originated in Europe. The ones that remain European are the ones that cannot be bought so easily, like Matrix (UK) or Mastodon (DE).

That makes the original point. They are bought by US companies because the US has facilitated tech and capital growth. In another world, the EU would've done better on that and bought US companies.

We certainly need more regulation to prevent this kind of buying. But that is clearly not the mood at the moment.

Is capital growth the most important goal we should set for ourselves as a society?

An increase in productive capacity seems like a positive to me.

Yes. That would've served Europe better too. Now they are in the embarrassing position of trying to regulate companies that aren't under their jurisdiction.

There are reasons to live that don't include being exploited for profit.

So you’d rather be more poor and less productive (but still exploited) just so that corporations would have lower profits?

I'd rather not be chained to serving humans at all, thanks.

Are you a robot?

The EU taxes profits, and economic growth results in increased employment and the tax revenue from taxing wages. This contributes to tax revenue and hence things like healthcare, social welfare, etc.

I don't necessarily agree with Singapore's approach, but Lee Kwan Yew made a speech that sums it up nicely...

"The opposition [in Parliament] say we're obsess with profit. I said 'Of course! That's how Singapore survives!' What do you think pays for all this? [talking about public housing, healthcare, public education, etc] You make profit into a dirty word and Singapore dies."

None of these companies except Nokia and Skype were even bought out by US companies. And those two are not exactly huge losses for Europe. The entire comment is BS.

The US doesn't have that much of a lead when it comes to hardware, especially chips, and is as dependent on Europe as Europe is dependent on the US in turn. Both are dependent on Taiwan anyways.

The US has a lead on software, mostly due to an inherent advantage of having a large domestic market with mostly the same rules everywhere, whereas the European market is as big in total, but unfortunately is divided into many smaller markets with fairly large differences and cannot be captured as quickly as the US market. This gives companies which scale nearly effortlessly from a tech perspective (modern software based startups) a huge advantage in terms of the available headroom for quick, initial growth. This advantage is largely irrelevant in industries where physical stuff must be produced and which are therefore hard to scale anyway, but it is almost a guarantee for a dominant role in the realm of SAAS.

So, no - there is no other world where Europe would have bought out US companies, at least not if we talk about Google, Salesforce, Oracle, etc., and if we talk about a similarly-structured US and EU.

Isn't ASML a Dutch company? I can't seem to find any evidence they've ever been sold.

Neither Spotify, Arm, or ASML have been bought by US companies.

This is easily shown by their Wikipedia pages.

Arm being the one that's currently listed on the nasdaq exchange and otherwise owned by softbank isn't a totally compelling example of a European company

Just because it lists in the US doesn’t mean it’s a US company.

It is also not a US company. Where it's listed doesn't matter much. It matters where a) headquarters and other significant physical presences are located and b) where the majority shareholders reside. In that regard it is largely a European, and partially a Japanese company.

You may want to review your info, these companies haven’t been bought out by US corporations

edit: I mean other than Nokia and Skype of course

Nokia hasnt even been bought by US companies. It sold a branch that was unsuccesfull and after it was bought it was still not successfull and they have the rights back. People are just ignorant about so many things and claim it is EU regulations that are the problem but that is a very simplistic view that just goes to show they are not interested in actually looking into what is going on and why

Ah yes I forgot how dependent the world is on checks notes Skype and Nokia

Do you know what Nokia does as a company nowadays? You might be surprised when you find out.

Surprise us

They are on the infrastructure side of cellular networks now, they are not irrelevant just because they're not seen by the consumer.

And the Nokia rubber boot factory is still doing well, after splitting from the main company in 1990.

Nokia is a key player of backbone communication tech.

Before the advent of iPhone and Android, Nokia was a really heavyweight player in mobile devices space.

But nobody is guaranteed their place if they make a strategic mistake. Ask GMC, Chrysler, Ford, etc why, when I'm going down a street, I see that more than half of the cars are Japanese.

We got BAE Systems and QinetiQ so...

Neither is EU.

UK, not EU but then the OP mentioned European Citizens, I'd argue I'm still European, that's geography.

> Is nobody in the EU bringing up the question why every single European citizen is dependent on US hardware, US software and US services?

And even more concerning is the dependence on US aid and weapons for Ukraine. How Europe doesn’t see Ukraine’s invasion as an existential threat is beyond me.

If isolationists like Trump come to power in the US and do things like pull out of NATO, Europe might be suddenly in serious trouble.

This is not a far fetched possibility. Even if Trump loses or isn’t even on the ticket, an increasing number of younger Americans are opposed to overseas involvements including some on the left side of the spectrum. Ballooning US national debt interest and social security issues could force the US to make steep cuts in places like military and overseas aid.

You need to remind those americans that Europe buys a lot of military crap from USA, pulling from NATO might cost USA more, and frankly Europe had a lot to pay from USA operations around the world by sending peace keeprs there or by having to handle the refuges.

I would be that pulling from NATO will not make those military money be moved into healthcare, I heard this story with Brexit.

European nations might buy a lot of hardware from the States, but the US also maintains a large and very expensive military presence there, and everywhere.

So if military relations cool down, the US pulls out of Europe, and Europe exclusively buys its own arms, it might actually save the US money.

But that seems a little far-fetched. There's a lot of transatlantic banter on the internet but when the chips are down, the West will likely pull together yet again.

Culling American military bases overseas would also reduce the risk of American politicians capriciously choosing to use those military bases as the staging points for military adventurism; the "War on Terror" was greatly facilitated by those bases. As somebody who thinks that 'war' was a disaster for the interests of average Americans, I would not be sad to see the means for America to wage foreign wars greatly diminished. I simply don't trust the American political system to be a responsible steward of that much global military power.

You may be right, but who else would do it? I feel like Russia would be invading the entire EU, and China would invade Taiwan if they thought they could get away with it.

Countries should be left to handle their own affairs; that's the responsibility that is inherent to sovereignty. The premise of a global hegemon who babysits the world and enforces the rules (dictated by themselves and their allies of course) is an anachronistic leftover from the age of imperialism. If the rest of the world wants America to play world police (and for the most part, I don't believe they do) let them pay the American public for this service.

Very reasonable, but the last time an American president suggested this, the world shouted him down as an idiot.

In a world where we have 9 nuclear powers, I struggle to see how your system would keep them in check; nuclear fallout effects all of humanity and quite literally could make the planet inhabitable.

I bring this up because any system that replaces the current world order needs to be able to keep these nuclear powers in check.

I wouldn’t bet on that last sentence. The leaders were about 100x more mature then than the ones we will likely have at the time we are next faced with an existential threat (meaning 1-10 years from now likely). Everything is a meme and a joke and a way to score political points. If a politician thinks they can own the libs / stick it to the filthy republican right by causing/allowing the global system collapse into complete chaos, they’ll happily do it.

Now, they’ll regret it instantly as they suddenly notice how everything they take for granted is suddenly not true (imagine if, thanks to abdicating naval superiority, sea transport becomes nearly impossible thanks to rampant piracy). But it’s not guaranteed that things can be undone.

> You need to remind those americans that Europe buys a lot of military crap from USA, pulling from NATO might cost USA more, and frankly Europe had a lot to pay from USA operations around the world by sending peace keeprs there or by having to handle the refuges.

Since when do populists make decisions based on logic and long term outcomes? It wouldn’t be weird for an American leader like Trump to act irrationally, he doesn’t even need support from most Americans to execute his whimsy.

Seems that we are assuming that net number of refugees is greater because of US action. I have no idea how this calculus is done though.

If the US wants to end Pax Americana, a strategy that has served the US extremely well, that’s on them in the end - it’s their choice. They shouldn’t be surprised though when it leads to losing out on a lot of trade, loss of influence, etc.

You must live in a bubble if you think that EU leaders don't see the invasion of Ukraine as an existential threat... Do you get all your news about what happens in Europe from HN?

Not OP but just going off the numbers, the US has been contributing more military aid than all other countries combined. I'm glad everyone is supporting them but a bunch of the European countries dragged their heels through the first several months of the war while the US was sending stuff by the ship load.

I get that the US is one big war machine but it's not like the EU doesn't have anything to give, they just decided that their own territorial defense was more important even though any supplies would be going towards fighting the only threatening aggressor in the region.

I think the problem is that EU countries don't have much military themselves to begin with.

They have a lot more than they're giving. Almost everything so far has been old models that were already replaced

> just going off the numbers, the US has been contributing more military aid than all other countries combined

The US has given the most in absolute numbers, but that's mostly a result of the fact that the US is by far the richest nation in absolute numbers.

When you look at aid given to Ukraine as a percent of GDP [1] you won't find the US even in the top 10. Instead the biggest contributors are European nations that directly border Russia.

Thus I don't think it's accurate to characterize the European countries as dragging their heels. That's more of a Germany thing.

[1] https://www.ifw-kiel.de/topics/war-against-ukraine/ukraine-s...

Most European countries have let defense production pretty much idle or stagnate for the past couple of decades under the blind assumption that things would remain stable, and if things went tits up, being allied with the US would fix everything.

During the GFC defense spending was pretty much the first thing to get axed - and it never recovered IMO in most places.

The leadership in some countries also has been frankly fucking delusional about the threat posed by Russia.

Now everyone’s getting that kick in the arse, a wake up call, and learning that expanding production is extremely difficult when you no longer have the factories or workforce capable of producing stuff like 155mm shells.

Well, yeah. Now they are. Obama was begging the EU to take this seriously back in 2014.

"Our people and our homeland face no direct threat from the invasion of Crimea. Our own borders are not threatened by Russia’s annexation. But that kind of casual indifference would ignore the lessons that are written in the cemeteries of this continent. " - https://www.youtube.com/watch?v=KbeLAQ3wBIs

I applaud Germany for weaning itself off Russian oil, but it's kinda weird that America seemed to be a lot more alert to this threat than the EU.

European here.

Ukraine/Russia war is caused by disagreements between US/NATO and Russia.

EU has nothing to do with it, and if US didn't involve itself with Europe or try to expand NATO, there wouldn't be any war. The chance of it spreading to the rest of Europe, except if EU countries attack Russia (or help Ukraine with too much weapons), is zero.

We in eastern europe want in NATO, our Western European would sell us for Putin's gases.

And he doesn’t realize that what keeps it from spreading to the rest of Europe is precisely NATO. So ironic.

What I’m reading here, honestly, is that mainstream Europe would readily abandon Poland, Ukraine, Finland and who knows whom else, if it meant not having to lift a finger against Putin. Not a good look my dude.

This is why Poland has been so supportive of Ukraine. They know this (one could say they are learning from previous experiences).

Are there predominately ethnically Russian regions in Germany or France on their border with Russia?

I think Berlin and Paris are fine for the foreseeable future.

World borders would look quite different if they were based on majority language. Would you redraw them all? Also, do you not consider anything east of Germany part of Europe? Wouldn’t you be worried if you were Finland or one of the Baltic states?

Also I find it hard to believe that Putin wouldn't come up with another excuse after that one ran out.. Did you see the https://en.wikipedia.org/wiki/Foundations_of_Geopolitics ?

> How Europe doesn’t see Ukraine’s invasion as an existential threat is beyond me.

Weren't we told that the whole war is about food exports from ukraine which euro countries like germany depend on.

Is the 'food is new oil' narrative not well accepted. Places like usa cannot export to germany because farmland and rivers have been poisoned by gylphosate

“Poisoned by glyphosate”

Fine starve then? I think most reasonable people would eat food from America over starving.

most ppl aren't germans. Germany banned glyphosate.

And irony of ironies, Bayer, a German company, purchased Monsanto and is currently stuck with the bill from all the people harmed by glyphosate.

Reality is stranger than fiction.

EU has very strong agriculture subsidies, it certainly does not rely on food imports for basic goods.

The whole European food mountain was a thing. And we're still wasting plenty of food, we're not going to starve any time soon.

How exactly do you see Russia as an existential threat to the EU? Putin does not seem interested in conquering the EU, at all. He’s never mentioned anything of the sort that I know of, and he’s not the guy to mince words about his intentions. He’s not even really interested in conquering all of Ukraine, because he did not at any time mobilize the kind of force that would be needed for such an endeavor.

I mean, yes, Russia’s nukes are an existential threat, but to the whole planet, not just the EU.

The EU doesn't aspire to economic or commercial success. It's an ideologically motivated socialist state with no real ideas for how to fund a high quality of life for the majority, but that's OK because the stated objective is equal quality of life, not high quality of life.

Except for the few people on the top of the hierarchy, they warrant maximum quality of life as is traditional in such power structures.

This is the vibe I got. For better or worse, it seems people in Europe enjoy/pursue equality more than individual success. I know many people that move to the US to "escape" that system though and they've been quite successful.

The EU isn’t a state at all. Getting tired of usually Americans talking about the EU as some kind of superstate. It’s nothing of the sort.

This is a European using terminology likely to resonate with the American readership. It's a government that spans and partially controls multiple countries. The analogy to the US government that spans and partially controls multiple states should be readily apparent.

You're using nonsense language to make an ideological point. You're blurring things out instead of clarifying them.

There is not such thing as an EU "state", period.

The EU is not in anyway socialist. It’s a neoliberal Capitalist federation, explicitly so in all its founding documents and philosophy.

Apart from the fact that the EU is not a "state", the claim that it is in any way or shape "socialist" flies in the face of basically all available evidence (including the leadership of most EU countries, the composition of the EU parliament, the laws that are actually passed, ...), so it's hard to read your comment as anything but flamebait.

If I was going for flamebait I'd call it communist and expect an uphill battle to persuade people that's the right adjective.

I'm taken completely off guard that calling the EU socialist would be in any way contentious. What's all the stuff about consumer protection, free healthcare and equality for all characterised as if not socialism? How can the hostility to corporations (e.g. this opening post) be characterised as capitalist?

I suppose I'm going off behaviour as opposed to what they say about themselves. Maybe that's where the terminology has gone awry.

> I'm taken completely off guard that calling the EU socialist would be in any way contentious.

Well, then you should look up words you're using. Nobody considers the EU to be socialist. The word "socialism" is poorly defined, but as a minimum it does away with private ownership of the means of production. No such thing is happening in the EU - private corporations are well and alive.

Every first world country (and many others too) bar the USA has universal healthcare (it's not "free", we have to pay for it), and even the US has consumer protection (sometimes even more so than the EU) and laws that companies must obey. And of course, no place on earth has "equality for all", certainly not the EU, where wealth inequality is quite high.

You didn't actually read the announcement. The EU is only regulating "high risk" uses of AI - i.e. facial recognition as used by law enforcement. Your spicy autocomplete and image generators should be safe.

Also, the EU didn't "miss the Internet", it got muscled out of the Internet by American tech monopolies.

> the EU didn't "miss the Internet", it got muscled out of the Internet by American tech monopolies

This sounds like a distinction without a difference. The American tech monopolies only became monopolies because of their first mover advantage. The EU had no Internet companies with first mover advantage... in other words, they missed the Internet.

Monopoly building is not a function of first-mover advantage. In fact, being the first mover is usually a disadvantage, because you have to figure out how your business model is going to work. Late movers can just copy it but better.

Going down the list of US tech companies:

Google did not have first-mover advantage. They were actually very late in search. They won search by having a superior algorithm and, more importantly, enough business discipline to not deliberately ruin it. More importantly, with the exception of Gmail, every other successful arm of Google has been a startup that they bought and scaled up. In some cases, notably YouTube, they were even competing with the company they'd later buy because they couldn't build the business themselves.

Apple did technically have first-mover advantage, given that they've been around in personal computing since the late 70s, but they blew it so hard they nearly went bankrupt. The Apple we know about today was destroyed and rebuilt by NeXT and Steve Jobs buying Apple with Apple's money. Even then, a good chunk of their success has been due to their purchase of several chip and core design firms that give the iPhone an edge that competing phone manufacturers can't buy. Even the first iPod was designed by someone else and bought by Apple.

Microsoft... also wasn't a first-mover. OK, they were there at the start of PCs selling BASIC interpreters, but the IBM PC DOS you probably are thinking of as their "first OS" was SCP[0] QDOS, which they bought and customized for late-moving IBM. The same story goes for the Internet - they saw Netscape Navigator, got scared, bought NCSA Mosaic, rebranded it to Internet Explorer, added a bunch of nonstandard extensions, and integrated it deeply into Windows to kill Netscape dead.

Is there a pattern here? Yes - that monopolies form because companies can buy instead of build. Why this happened in the US and not the EU is simple: the US stopped enforcing its antitrust laws and gave tech companies a regulatory pass in the name of the "future". This is not the EU missing the Internet, this is the EU being the victim of cheating.

[0] Seattle Computer Products, which I'm assured is not a front of the SCP Foundation

Of the companies you claim weren't first-movers, your counterexamples of the "real" first-movers were also US companies. This would be a more relevant argument if some EU companies were first-movers but then got "muscled out" by US late-movers that became monopolies (and an even better argument if the EU companies were impeded by EU anti-trust rulings against them). But that's not what happened, because there were no EU first-movers, nor were there any anti-trust enforcements against any would-be EU first-movers... It's like saying you lost a race because your opponents cheated, when you weren't even at the starting line.

> the US stopped enforcing its antitrust laws and gave tech companies a regulatory pass

So the EU did enforce its antitrust laws during that time, and thus stopped any EU tech companies emerging as monopolies? Can you provide an example? (In fact the only anti-trust ruling I remember from early Internet era was from the US, against Microsoft.)

Besides, has US ever considered anti-trust law in the context of the international market? Do they even have an obligation to do that? The government has a responsibility to its own citizens. And if what you say is true, that US neglecting to enforce anti-trust law facilitated US tech monopolies, then perhaps US citizens benefitted in the long term. Certainly I'd rather be a citizen of the country that's the dominant economic power than the subjugated one in decline... and it's not like those tech monopolies halted innovation in the US; on the contrary, it's thriving. The US has the highest levels of high-tech investment, the highest salaries, the latest tech...

I would agree with this comment, except for the last part. Antitrust law enforcement is not the critical part. Uniform market size is.

The US has a lead on software mostly due to an inherent advantage of having a large domestic market with mostly the same language, rules and expectations everywhere, whereas the European market is about as big in total, but unfortunately is divided into many smaller markets with fairly large differences which cannot be captured as quickly as the US market. This gives companies which scale nearly effortlessly from a tech perspective (modern software based startups) a huge advantage in terms of the available headroom for quick, initial growth. This advantage is largely irrelevant in industries where physical stuff must be produced and which are therefore hard to scale anyway, but it is almost a guarantee for a dominant role in the realm of SAAS.

The remaining conclusion of yours is true: if you grow faster, you can buy. If you grow slower, you are being bought.

Interesting, I never really considered the relative homogeny component but one I find convincing as a notable advantage.

That said, I'm not convinced that it is more notable than Antitrust law enforcement. Scalability is great but it market dominance is king; if your able to purchase any potential competition or drive them out of business with relative regulatory impunity then you can just buy marketshare and worry about scaling after. (see amazon)

Absolutely. I think we are not going to see the end of that inflation anytime soon because of that. We have nothing to put in front of Google, Apple, Microsoft. Nothing.

The issue though is we need to federate. We are not Europe yet really but a bunch of small countries. The political setup is not there for big tech companies to thrives.

federated but united... we kinda already are, no? see all the state level privacy laws

> US hardware

Chinese hardware.

> Europe missed the internet

What the EU doesn't have is large scale social networks, and large scale "infrastructure" companies (the likes of Google and AWS) but it is missing less than what your average US perspective sees. No wonder most US companies sell a lot to Europe (including B2B).

> is dependent on US hardware, US software and US services

True. And the US (and the rest of the world) is dependent on HW produced on TSMC and assembled in China. It truly sucks

Which in turn relies on litography equipment made in the EU and Japan.

Disregarding the regulation itself - the logic you describe here is that non-EU competition isn't regulated, so EU shouldn't regulate anything to be competitive.

1- the EU didn't "miss the internet" because of regulatory policies. At least not regulations on the Internet.

2- even if that was the case, does it mean the EU should always level itself by the worst competitors?

3- and then, assuming it deregulated LLM, how would would it lead to a place where EU would suddenly becomes competitive on that market?

It missed the internet in the sense that it is near impossible to create an innovative (read: disruptive) internet company within the context of EU and member state regulations. So people don’t bother and entrepreneurs emigrate to the USA.

It looks like the EU is preparing to miss AI as well.

There are a lot of other countries that are not EU which also have minimal big internet companies. Did all of them regulate badly too?

I suspect its more the power of consolidation and first mover advantage.

And having a common language across the market.

I could honestly do with less disruption by internet companies.

Europe “missed the Internet” (and the US didn’t) because of various underlying factors - smaller/more fragmented markets, a more muted entrepreneurial culture, shallower pools of capital, etc.

I don’t think the fact that Google, Amazon, and Facebook were built in the US is down to EU regulations.

It’s no more difficult to start a tech company in Europe than it is in the US, and it’s generally no more onerous to grow it, from a regulatory perspective. But ambitious entrepreneurs often head to the US in order to tap into a larger market and more developed ecosystem.

So I’m not sure the analogy is that valid. But yeah, I do think that EU regulation of AI may be an additional factor that skews the playing field.

> shallower pools of capital

Why do you think all the capital is allocated in the US? Because we don't smother as many companies to death

> It’s no more difficult to start a tech company in Europe than it is in the US, and it’s generally no more onerous to grow it,

This is an extremely bizarre claim.

> I don’t think the fact that Google, Amazon, and Facebook were built in the US i

I'd say that even more importantly, they were born in very specific places in the US

It's not like Amazon could have started in Atlanta for example.

(Amazon does sound like the one that was the most distant from the academic and valley world, but MS was in Seattle first)

I think there is more to it. The USA has well the U and S - a homogenous enough and high population bunch of wealthy (relatively) English speaking states with similar enough laws/rules when it comes to commerce that you immediately have a large incubate market before your product can go global. This probably network effects into more funding, which creates a feedback loop, centered on the schilling point of Silicon Valley / SF etc.

If I needed a $1m for an idea, I can go beg for it in Australia, UK etc. have the Dragons Den types laugh at me for not giving them 99% for $100k, or I go to the US and have a different problem - how to choose which investors to dedicate my time to, and which are wasting my time. But getting the money is not the problem.

> It missed the internet in the sense that it is near impossible to create an innovative (read: disruptive) internet company within the context of EU and member state regulations

Can you elaborate on that? And was this also the case in the 2000s and early 2010s?

Popped up on HN today. https://www.unite.ai/paris-based-startup-and-openai-competit...

There is no fair use for copyright in Europe. That's the reason why we don't have any google here, and won't have any AI company here.

At least in the UK we do - https://www.gov.uk/guidance/exceptions-to-copyright

DeepMind was also founded here, 5 years before OpenAI was.

Some level of regulation clearly is necessary, but it's also clear that it's possible to over-regulate (or mis-regulate) to the point that your regulations become meaningless because the activity you're attempting to regulate simply stops happening due to the added burden.

It's sometimes a bit hard to tell whether the EU's goal is to regulate tech in a way that still promotes its development and creates a net good for society or to just stop most development of new tech within the EU wholesale.

I've strongly considered moving from the US to the EU (Spain) since there is a lot to love about the culture, but the bureaucracy and arbitrary/capricious nature of regulations make it hard to justify trying to run any kind of innovative business there unless you have enough funding to pay a team of lawyers to protect you.

As a spanish I tell you that we are drowning in regulations. Not only european, but Spain itself also.

The level of taxing is also too high for the salary ranges and cost of life.

It is a pitty, because I really think my country has more potential but it is terribly managed from the POV of economics. It could be way better :(

Having visited your country, your lifestyle and quality of living is so much better than a “middle class and below” standard of living in the US. I don’t think you want to give up what you have for what we have unless you’re pretty rich.

But we have a big debt. Health care system is not what it used to be. I think we are living a fiction and this will end up in less than a decade.

We will still be spanish for our spanish things as much as we can but... I am alert of what could be coming.

I got that feeling in Europe in general. I had no idea how these countries were paying for all the nice things they have. I just assumed it was some long tail consequence of colonialism that keeps bringing in revenue somehow or “money makes money” from the massive wealth transfer during the colonial period. On the flip side, healthcare can bankrupt you with one semi serious illness in the US and there is no work life balance.

Colonialism does not make countries rich. That is amyth with data in hand. It makes a few people very rich but not a country as a whole. In fact they historically brought deficit in many cases. Even if looting happens. I tell you this as a spanish that checked the topic.

I do not know about USA lifestyle enough and I know americans work hard. There is no trick here though. Without enough productivity and generated wealth, it is impossible to maintain a system forever. After living for 12 years almost in Vietnam, which is a country more similar to how most countries are wealth-wise and conditions in those countries are harder.

I reached the conclusion that spanish demand too good conditions compared to productivity for a lot of jobs. This will eventually make us collapse. Same for some parts of Europe. I agree we all want to live better... but careful with balancing it towards our own personal wishes. Because we could eventually regret it :)

With respect, Spain is one of the cheaper places to live in the EU. Perhaps your salary is even lower, but the nice thing about the EU is that you are free to get something better in a place where you get a salary that's appropriate for you. If that's not an option, you must live in a cheaper location. That really has very little bearing on your country's governance.

No, my salary was was above average. But the taxing starting at 62000 EUR, which is not that much if you want to buy a house and have a family is at 47%. The higher you try to go the more they smash you. The PIT at 300,000 (which is a lot, but anyways, you already pay more at that rate) reached 54%. After that you have to add things such as IBI (tax for owning a house), indirect taxes such as VAT (mostly 21%), gas (50%), electricity (50/60%), garbage collection, if you have a car you pay 7 taxes more... freelance workers pay around 300 euros just for working, every month, before having any incomings.

After that they pay VAT and social security. If you see the average salaries in Spain you will understand that what they are doing now to the freelance workers (autónomos) is an abuse: a person that does not know how the coming months are going to be gets taxed increasingly more over the years and they did a reform that will raise the costs for this group (depending on earnings) by a lot. It can even quadruplicate the costs per month for people that get around 5,000/6,000 gross.

If you start to count and they can smash you like 60% or more. Things are going to go higher on top of that. It is abusive.

With that amount of taxes the debt should be lower and the services better. Where are we spending the money? Mostly in pensions (9.5 million people in Spain).

In Spain you have young engineers working for less than 1,500 eur/month, some working for around 2,000 (I mean Master's degree both), but you have local police, which is two years studying, which get 1,800 eur/month. Some people have pensions of over 2,000 euros at the same time, because the pensions are a Ponzi scheme. We have as much as 3.5 million public workers. To get an idea, in Germany, with 83 million people (vs. 47 million in Spain), they have around 800,000.

There is a lot to fix... and the unions... do not get me started on the unions and politicians. In the meantime, the private sector is going weaker and weaker and paying higher and higher bills to maintain all this.

it did miss the internet, and partly thanks to uncompetitive regulation. You will not find the equivalent of gdpr that targets tourism or car manufacturing or fashion because those are large moneymakers in europe. Conversely, no large subsidies for internet or AI either. These industries don't just appear spontaneously

> 3.

MistralAI is a european company. It makes an unregulated open source AI and they are proud of both , it s actually part of their mission. It worked and it s becoming increasingly popular and successful. How will mistral grow to the size of openAI etc?

>car manufacturing

Are you serious? EU regulates cars much. Majority of cars sold in U.S. cannot be imported to EU due to them failing safety regulations.

>cars sold in U.S. cannot be imported to EU due to them failing safety regulations.

This is not a bad thing.

Living in the third world, there are "exclusive" cars here from big manufacturers too; but they aren't available in EU because they're 0/5 safety rated tin cans. I keep an eye out if a car is "only" sold in the global south and avoid it like the plague.

I have heard of the reverse occurring as well, EU cars which need modification to be road legal and not from "steering wheel on the wrong side" issues. I believe one example were smartcars ironically failing emissions tests. The point being incompatibility may be mutual.

The opposite is actually true, most likely the lack of regulation is why we don't have more European hardware and software. Look at Russia and China - they have much more onerous regulation, and have large popular local services (VKontakte, WeeChat). I'm not suggesting in any way we should want to be like China and Russia, but I think they prove that regulation doesn't prevent local business springing up to compete with international behemoths.

Those countries simply forbid competitors from selling in their domestic markets.

EU regulations apply equally to EU and non-EU companies, so they won’t lead to homegrown tech champions. More likely they will cement the dominance of US tech giants because they’re the ones who have the resources to follow these regulations.

Facebook was available in Russia up until the invasion of Ukraine. China doesn't directly block most such companies either, it imposes censorship and technology control requirements. Microsoft, Google, and Amazon operate in China, with Chinese subsidiaries, but they are nowhere near as dominant as they are in the rest of the world.

Not exactly true, up until 2022 Google, Meta and others operated in Russia and were competing with Yandex and VK on a more or less even playing field

>popular local services

>Vkontakte, WeeChat

Since 2008 big US tech companies have been a runaway GDP doubling success. It is laughable to compare them with VK and WeChat.

The reason we Europeans should aim to have our own big tech companies is because we are currently missing out on the greatest wealth generation in history.

> because we are currently missing out on the greatest wealth generation in history.

You mean losing billions dollars a year, for years, with no plan of becoming profitable? That's what most of "successful startups" are in the US. It's wealth redistribution for a lucky few, and gig economy for everyone else.

I think you misunderstand how EU works. In reality it s a social club where mediocre politicians become careerists and climb the social ladder with a practically unlimitd budget of german money. There is no dialogue with any society other than with the brussels bubble and sometimes european academia. Some ambitious people latch onto the popular matters in the media like privacy, AI, climate, blockchain etc, and naturally they want to become a part of it by creating ... rules.

There was a time decades ago when EU seemed to have some cohesion, nowadays it s just ambitious individuals who want power for a few years and then go on with their lives to other endeavors. It's almost all process and a lot of marketing-speak these days. Really a lot of marketing-speak , that sounds frighteningly more and more like the ESG and DEI departments of corporates. It s become soulless

I often ask myself that too. The talent and infrastructure definitely exist in Europe. (until the people leave for the US) Is it the big VC money, or some magical governmental support (military-industrial complex)? is it the english language that gives the US-companies an advantage? is it the different mindset? I really do not know.

It's very weird and also sad and a bit frustrating to see that the market is in favor of one country that seems to get everything. I'm still crying for Nokia.

Yep, I've been wondering this too. It could be:

- America has a large unified market, it's easier to expand than in the EU.

- Immigration, smart and ambitious people move to America so their companies have access to better pool of talent. Better labor mobility compared to the EU also helps.

- Cultural differences related to entrepreneurship and work.

- Europeans who have historically moved to America have specific genetics and learned behavior, which is passed over generations. Maybe they're less risk-averse, more optimistic or have more energy and desire to do something. Basically, if you select a subpopulation which is willing to move to a different country hoping for a better life, this subpopulation will have distinct traits.

- Legislation, taxes?

Europe should first identify this as a problem and then try hard to understand and solve it.

I remember not long ago during the large tech layoffs, people in France and Germany were proud that the regulations meant they couldn’t be laid off, and looked down at other primitive countries without such regulations.

I wonder how attractive it is for Google/Microsoft/etc to start new teams and hire engineers in those countries. Probably not so much because of the risk.

I also wonder how these regulations affect salaries. For the same cost I would much rather hire an equally talented US engineer and avoid regulations holding me down. Therefore EU engineers must be cheaper to hire.

Maybe EU citizens want to trade things like job security for fewer jobs and lower wages, but I feel like people mostly imagine regulation is consequence-free.

The difference in salaries is sooooo vast that it make sense.

A new grad at AWS makes about 45k gross in Poland, 55k in Germany. That’s about 1/3 what a junior makes in the US. 90-110k for L5 in Germany, 80k in Poland, and 250k-300k in US.

It’s a no brainer all things considered.

how do yoh mean europe missed the internet? my experience since i left the us years ago is that the tech used in eu is leagues above that from us companies and often feature equivalent or even better. banking and payments are a fine example of this as instant transfers were built into the banking system as i get it for some time and every bank i’ve used has had instant transfers by phone, card, or iban for years. even when i was in russia it was pretty common for pop up shops on the street to accept either cash or direct payments via a phone number they posted. even quite a few shops used this model.

i would not call the tech for common US life particularly good or something that is desirable to emulate. i thjnk europe is doing pretty well wrt the internet. it’s no coincidence for me that EU sites usually follow gdpr as its written and supposed to be and include a clear “deny all” button typically whereas many us and uk sites i try to read have pages of obtuse buttons and menus to manually deselect dozens (sometimes hundreds) of tracking “features” on their sites. (not 100% every site ofc but it’s my general experience)

i am not worried about being late to the party with ai either as the current offerings from the big players either are neutered to a pretty high amount or the companies are unabashed regarding their intent to serve more fuckjng ads via ai. i’ll gladly pass on that until we see a better alternative as the benefit for me from ai is minimal, and at best so far for my line of work ai just adds noise to already noisy conversations and more things to research and verify/fix than it does help.

i cannot see that the current ai hype is at a good and useful stage where i want to pay for it for daily use. it’s great and interesting tech for sure but it’s not worth it for me yet. that the ai based game where the goal is to trick the ai to give you passwords has been the most useful ai project i’ve used speaks a lot to me, as the “bad” and unexpected results are part of the challenge and fun of the game.

as pointed out by others, this is a myopic view. as much as "AI" has been overhyped, most of the regulations do not prevent the advancements in the field.

doing it right seems better than the startup grindset of experimenting and failing fast imho.

if you look away from consumer (and generic-use business) space, a lot of software and hardware has local players. to think you can do the whole stack inside europe is a pipe dream, however. even if you go wild west on the regulations.

https://twitter.com/ylecun/status/1733481002234679685

> The EU AI Act negotiations ended. One contentious issue was the regulation of foundation models, particularly open source ones.

> Kudos to the French, German, and Italian governments for not giving up on open source models.

> Juicy part: "The legislation ultimately included restrictions for foundation models but gave broad exemptions to “open-source models,” which are developed using code that’s freely available for developers to alter for their own products and tools. The move could benefit open-source AI companies in Europe that lobbied against the law, including France’s Mistral and Germany’s Aleph Alpha, as well as Meta, which released the open-source model LLaMA."

> which released the open-source model LLaMA

While this part is inaccurate, the open-source stipulations of the regulation is something I applaud too, as it would be too easy to have regulatory capture by the big companies such that only they could afford the licenses necessary to create AI and thereby have onerous influence over our lives. More software, especially of the critical variety like AI, should be open source.

The EU has provided a high level overview of the regulatory framework here: [1]. Quite old (2021) but the general idea is probably still valid. I think the idea is sound.

The question is whether foundation models will be regulated already as high risk because they could be applied in high risk use cases or if only the use case applications themselves will be regulated.

It tends to be more in the direction of "let's regulate the foundation, as well".

BTW: Something similar is in the works for regulating software, including open source software with the new "EU Directive for Liability of Defective Products". Here is a discussion: [2].

The rationale behind both seem to be something like: We (software vendors in the EU) rely so heavily on open source software and put just a little bit on top, that it would be too much work for us to ensure compliance of the underlying software for our use case. So, please, dear regulators, regulate the underlying software, as well, even if it is open source, so that the compliance burden is not on us, but on the authors of the underlying software.

The compliance industry is happy (new business), the big software vendors are happy (reduced competition from smaller vendors in high-stakes industries), the regulators are happy (fewer vendors to deal with while ensuring the regulation will actually be enforced).

/sigh

[1]: https://digital-strategy.ec.europa.eu/en/policies/regulatory...

[2]: https://www.taylorwessing.com/en/interface/2023/open-source-...

https://www.theverge.com/2023/12/8/23991850/eu-ai-act-artifi...

https://artificialintelligenceact.eu/wp-content/uploads/2022...

I was reading through this a couple of evenings ago to try and get up to speed.

I find the spirit of the law to be something I broadly approve of -- for example: "There are a number of applications where the use of AI is banned, like scraping facial images from CCTV footage, categorization based on “sensitive characteristics” like race, sexual orientation, religion, or political beliefs, emotion recognition at work or school, or the creation of “social scoring” systems." These things seem broadly good to keep at bay.

However, laws are more than just what they intend - it's what they actually do, how they're enforced, their secondary effects, etc. People don't like being tracked without consent, but people haaate all the "This site uses cookies" popups.

From that perspective, I am not so sure about all of this. The language feels like its grasping at straws a bit prematurely in an attempt to head-off a problem which isn't fully manifest yet. Yes, it's good to get ahead of this sort of thing, but, IDK, I just wasn't fully convinced the legalese knew what it was on about. It even admits as such, stating that it encourages using broad terminology to future-proof itself and encompass the rapid pace of innovation.

Plus, yeah, as others have mentioned -- I feel EU is starting to litigate its way out of technical relevancy. It's one thing if the forefront of innovation were occurring there, and they were then dictating how that innovation should evolve, but that doesn't seem to be what's occurring. Technological progress is a cornerstone of strong GDPs and I'm starting to fear for the EU a bit in this regard.

> People don't like being tracked without consent, but people haaate all the "This site uses cookies" popups.

I thought that was the intent… surfacing ubiquitous tracking to direct consumer frustration at the practice itself.

I don't think so, because that consumer frustration has resulted in nothing. Nada.

We click the popups and get annoyed at the EU, not at the website.

Eh true. I wonder how that is different than, say, the tracking notifications that Apple added to iOS. Consumer choice maybe?

I guess they can also enforce a good UX by controlling what is listed/installable on the play store, so apps can’t maliciously comply.

Regulators are being slow, but with GDPR you can also not maliciously comply. You can get fined for that and slowly regulators are starting to so that

I always get annoyed at the website.

To the point where I'll click away if they don't give me a way to opt out, or if I feel like they're taking the mickey too much. All those cookery websites that are so snooty about cookies as if there weren't 10 more similar websites where they came from...

I really don't, I definitely get annoyed at the website, and I get way more annoyed at websites which make it hard to say no to the cookies easily.

Nada? It's made huge changes in how businesses use ad tech. And yes, I work as a martech consultant.

Pre-2018 nobody cared about sharing data with third parties. Today nobody wants to be this company. https://www.nrk.no/kultur/nettapotek-delte-med-facebook-at-d...

Ymmv, but nada is categorically wrong.

This is the same argument used in America for why taxes and expected gratuity aren't included in the item price, and it has done nothing to fix the problem.

I’m not 100% sure I follow. Are you saying this is done by retailers to dissociate themselves from sales tax and make it obvious to consumers which part of the total price is the governments fault, or something else?

Loosely, yes. The historical reason is a lack of state sales tax for many states in the 20th century, however one of the main justifications for continuing this process is to highlight what portion of the price is sales tax. However, it has become incredibly evident that large chains abuse this for psychological reasons.

In the EU and Japan for example, prices include tax, allowing for a more honest depiction of total cost to the consumer.

> in an attempt to head-off a problem which isn't fully manifest yet.

It has already manifested. All the dangerous uses of AI etc. are already there and used all the time all across the globe from China to the US.

> It's one thing if the forefront of innovation were occurring there, and they were then dictating how that innovation should evolve

Doesn't the AI act specifically says: document your foundational models, and gives preference to opensource? It's a much better approach than the black box of OpenAI.

----

> People don't like being tracked without consent, but people haaate all the "This site uses cookies" popups.

And the law that mentions neither browsers nor cookies is responsible for this how exactly?

That will be the main issue with EU AI Act, too: the industry will gaslight it, spread enormous amounts of FUD, will engage in years-long malicious compliance.

This. Making tracking annoying to consumers is feature, not a bug.

> All the dangerous uses of AI etc. are already there and used all the time all across the globe from China to the US.

If this is as bad as it gets then it wasn't so bad.

> If this is as bad as it gets then it wasn't so bad.

Mass surveillance? Social scoring based on 24/7 monitoring of your activities? Stories like this: https://www.politico.eu/article/dutch-scandal-serves-as-a-wa...?

You think this isn't bad? Considering we've just started all this?

> You think this isn't bad?

Remember I said "wasn't so bad". Compared to the speculations such as all jobs being eliminated and flawless AI impersonations leading to a total breakdown of communication, yes. It's bad, but it could be a lot worse. And governments can cause fantastic amounts of greater harm just by declaring war. No algorithms required.

> Considering we've just started all this?

Remember I said "if this is as bad as it gets"

> It's bad, but it could be a lot worse

Yes, and it's getting worse

> Remember I said "if this is as bad as it gets"

1. How isn't this already bad enough for you?

2. How are you blind to the fact that it is getting worse? With companies like Deloitte already pushing for things like predictive policing [1] among many, many examples of systems already in place, and already causing harm.

Yes, governments can go to war, but that's reductio ad absurdum. Whatever bad things happen you can always find worse things. Doesn't mean this somehow justifies or nullifies the existing situation.

https://www2.deloitte.com/fi/fi/pages/public-sector/articles...

I've heard the phrase "regulating yourself out of the market."

Do you feel this applies here?

Of course. I mean this is the eu we're talking about.

Happy not to conduct business with a party if its practice does not reflect my values.

So the answer is to take that choice away from you?

The answer is that unbridled capitalism is not what I want.

What about the people in Europe who do want to use those services? Why does what you want trump what they want, especially when them doing what they want has no direct impact on you.

If those other people start using tech to automatically recognize my face on CCTV footage, that absolutely has an impact on me. I'm using the CCTV footage example because it's one of the things this regulation forbids.

We don't have anything anywhere close to unbridled capitalism.

Are you ready to opt out to cancer treatments found by US AIs?

This sort of comment is so useless.

Do you actually believe cancer treatments are what the EU is going to lose on with these regulations?

Why?

While I agree that the comment you replied to seems a bit over the top, I don’t think the answer to your question is as obvious as you make it out to be.

1. If the AI is more effective at finding successful cancer treatment for Little Endians versus Big Endians, then there is a non-zero chance that the AI will be deemed discriminatory. This could play out in a number of ways, but restricted access is certainly one of them. I live in California, and I also think that there is a non-zero chance of this happening here. It all depends on how the treatment and data are presented.

2. If the manufacturer of said treatment is anti-union, will the treatment options be distributable in certain parts of Europe?

There were plenty of top tier Covid vaccines created in Europe, and a Danish pharmaceutical is currently crushing the US weight loss market: your jingoism is deluding you.

> your jingoism is deluding you

As is yours.

Who said anything about where I think this AI would be developed?

US is possible. So is China. So is Europe. Not necessarily in that order. Japan might have an outside chance if certain things line up a certain way.

If Europe and the US make AI research overly cumbersome, I’m sure there are plenty of countries that will offer sweetheart deals to the best teams looking for a new home.

It's not AI, but drug makers have already opted out of EU countries because of their regulations.

"Drugmakers Boehringer-Ingelheim and Eli Lilly have called off plans to market their Type 2 diabetes drug linagliptin in Germany because new legislation in the country could mean that pill's price could end up being too low."

https://www.reuters.com/article/boehringer-lilly-idINL5E7K23...

Indeed, I will have to geoblock my products out of the EU, as I've heard others have already done.

I'll go further. I fully do not recognize certain internet laws enacted recently by various states in the US, such as requiring identification to view a website, or banning children from using the internet after dark.

Laws like these are unconstitutional and I intend to actively disregard them while still serving customers in those states.

How are they unconstitutional?

First amendment grounds.

https://www.washingtonpost.com/technology/2023/09/18/califor...

Know your rights.

The axiom is that the EU is such an important market that it can dictate terms to the world. Such as the landmark success of persuading Apple that it may as well use USB on mobile phones.

It's not as clear that it's better to give up on artificial intelligence than to give up on Europe. When your axioms are unsound, you can derive nonsense from them. So yep, it seems Europe isn't going to be the future of machine intelligence, to the astonishment of noone.

> The axiom is that the EU is such an important market that it can dictate terms to the world.

So the EU exists strictly in an effort to bully the rest of the world into their continental ideas? I thought the purpose was to provide representation and protection to their own citizens.

> So yep, it seems Europe isn't going to be the future of machine intelligence, to the astonishment of noone.

They mistook State Authority as an end unto itself. Everything else is going to be secondary to that.

Before I sold my previous company, a medium size ai company, it was already often an internal debate whether the EU was worth selling to for our size company.

The reality is that the EU is a much much smaller market than North America, and expensive to operate in even without onerous AI regulation.

It’s likely easier to support the EU for larger companies, but there is little doubt in my mind that this sort of regulation in the near term will ensure that Europe will always get services much later than NA, and maybe some companies will choose not to bring some products to the EU at all.

The EU has been declining in its share of the world market year on year - it should focus on stopping the decline, not adding more regulatory burden.

What a trade. Sacrificed the next qualitative shift in economic growth after the industrial revolution, and in return was able to force the switch of the Lightning port in one type of phone for USB-C.

> Those that pose a systemic risk will need to sign onto a voluntary code of conduct ...

How can one need to sign onto a voluntary code of conduct?

Don't you have voluntary prison sentences in the USA? same thing i guess...

What's a voluntary prison sentence?

>Cognitive behavioral manipulation of people

from https://www.europarl.europa.eu/news/en/headlines/society/202...

Aren't the valuations of Google,Facebook et al mostly based on their future capabilities to do exactly this when they feed everyone's data into their AI?

The EU has a billboard campaign going recently with pro-EU messages and utopian imagery which is also be cognitive behavioural manipulation of people (manipulating their thoughts and behaviour (voting) to be more pro-EU).

Do you have any sources on that? I wouldn't be surprised though, with populism and Euro-skepticism at all time highs.

It's this one, "you are europe":

https://you-are-eu.europa.eu/index_en

The images and slogans in the 4 pictures are what I have also seen on billboards.

I saw one in Sweden recently and thought it was weird. Didn’t realise it was part of such a campaign. Feels totalitarian.

>Feels totalitarian.

Meh, it's just propaganda and not all propaganda is minted by totalitarian regimes. The US/west also minted their own during the cold war.

> The US/west also minted their own during the cold war.

And before and after, its not like the Cold War was in any way unique in seeing the US make propaganda.

Could you share some links about this?

See my comment to FirmwareBurner.

The EU doing a basic advertising campaign is somehow equivalent to mega corporation manipulating people’s behavior (to their own benefits) by controlling their content feed? Your comparison dilutes terms so much that the whole issue of algorithmic content feeds is lost.

Most advertising is manipulation.

*all

If advertising did not change behavior, it wouldn’t be a thing.

Content feed: text and images you see when you look at your computer/phone

Outdoor billboards: text and images you see when you go outside and look around

I don't see what is dilutive about this comparison, seems about equal to me (except outdoor advertising is worse because you can't avoid it by not visiting certain websites). My broader point, I guess, is that the intended interpretation of the word "manipulation" seems to be restricted to "manipulation we disagree with".

While billboards are annoying, a) everyone sees the same, they aren't tailored, b) they aren't as ubiquitous, most people don't spend as much time looking at billboards as they do at various feeds and c) they are only show 1 advert, I would say an equivalent to a feed would be every billboard in a part of a city showing an advert that sells the same thing in a slightly different way.

That quote is missing some context, the term is used in a section on “systems considered a threat to people”, the examples it gives are specifically behavioural manipulation to do something dangerous like violence. It’s not just any attempt to change someone’s mind on something, that would make all advertising illegal.

I agree the wording is too vague, though.

> that would make all advertising illegal.

I’d vote for someone pushing that.

As a bonus with all those American ad-driven malware driven out of the market there would definitely be room for European companies to compete on delivering actual value

Great. Glad to see we're still world leaders at shooting ourselves in the foot. It makes sense to regulate AI use in critical infrastructure and flat out outlaw use of AI to manipulate public opinion, but that's where it should have stopped.

AFAIK the actual final legislative text is not yet publicly available, so we do not know what was passed.

But might change any moment.

What?

How can legislation pass without the public knowing what is in it before it passes?

From what I understand, it's not passed yet. They've just reached a deal on the amendments that need to be added so it will be passed.

We do not know what amendments have been passed, as Germany and other countries were lobbying for more relaxed regulation after every AI company in the EU called them and told them the regulation was going to make them uncompetitive.

How many popup warnings will I need to bypass/accept in order to use my LLM of choice?

Just a brief conversation where you’d need to make your case to a friendly EU-hosted LLM

This all seems premature. So far we have something that can generate text and pictures based on other text and pictures ?

> This all seems premature. So far we have something that can generate text and pictures based on other text and pictures ?

I think the reasoning is that at one time Social Media was just a few sites where people could share photos and fun articles they'd read - and it turned into a privacy/hate speech/monopoly minefield that no one is sure is completely good for society and now is impossible to control.

why do you have to control anything? Again the government knowing no bounds, but not surprising coming from non-democratic bureaucratic EU

> why do you have to control anything? Again the government knowing no bounds

Because most countries have laws about discrimination, promoting hate, inciting hatred, meddling with elections etc.

There's nothing here that wasn't included as part of the Bletchley declaration about AI safety and signed by just about all the well known AI players including democratically elected governments last month.

Because not controlling anything has a LONG history of utter misery and might still lead to borderline human extinction as it is now.

Elites controlling everything is why the couple thousand years prior to the 1900s almost everybody in the world lived in conditions that would be considered abject poverty today.

Honey, the elites control most of the infrastructure related to this stuff now.

And you don't think conditions suck for a good number of the world?

This sort of rhetoric is empty and meaningless, useful only for antagonism.

All I'm saying is that's not an excuse for deregulation, it's literally a reason to do more regulation. Open source and distributed social media exists, that doesn't mean we should let meta do whatever it wants.

E: A lot of the libertarian rhetoric here falls apart when you examine other monopolistic situations like water or electricity. You need massive datacenters to run these applications. That gives the corporations who can operate that capital advantages that lead to them owning spheres of computing like AI. The only way to address that monopoly power is government action, unless you like the negative market externalities created by those monopolies (maybe you own equity in a technology company). Frankly, even if the EU gets it wrong the first time, I am much happier with their approach of trying something than I am with the United States approach of "we need to innovate at all costs! What if china catches up?!". I've had enough hysterics from free market people who think gigantic tech interests can do no wrong, or who are under the naive impression that the open source community can create these models without the backing of big capital.

And not having any control is why some "shithole" countries have that reputation..

Really? I know Snap AI (which is available for kids over 13) will happily advise you with how to cut yourself if you ask it.

So would Google or any obscure forums. Just like fraud, the optimal amount of negativity in a technology is never zero. Gotta accept some amount of issues to maximize the benefits to drawbacks ratio.

praying that the regulators do not read this: so far it can generate code and run code either in the sandbox or if you give it access to your system.

> This all seems premature. So far we have something that can generate text and pictures based on other text and pictures ?

People say this with a straight face having China as an example, and news like this https://www.404media.co/fusus-ai-cameras-took-over-town-amer... and Deloitte alraedy campaigning for predictive policing: https://www2.deloitte.com/fi/fi/pages/public-sector/articles...

They sure do pass a lot of new rules these days about sweeping.

its almost like they want to send all of their talent to the US and Japan

such a weird move

It's for fairness, the EU has to handicap itself after the US decided to ban abortion and persecute trans people.

Anyone have the text that was agreed upon?

Edit: I know the text is not final as it needs more work, but I was hoping for a leaked draft text of amendments that clinched 'the agreement'.

I don't think it's ready yet, from what I read in statements they found a compromise but it still needs to go through some processes to turn the agreed points into legalese (and then the translations)

I don't think it'll be voted on this month

This month is Christmas. It won't.

Yeah without the actual agreement (which is still subject to voting, etc) this is an empty discussion

This is literally based on one short tweet

Edit: this seems to be a (slightly) more substantial press release https://www.europarl.europa.eu/news/en/press-room/20231206IP...

"prohibits ... social scoring based on social behaviour or personal characteristics;"

Does this outlaw Tinder?

> AI systems that manipulate human behaviour to circumvent their free will;

> AI used to exploit the vulnerabilities of people (due to their age, disability, social or economic situation).

So basically all social media in their current forms?

All that don't just have time ordered feeds, or feeds which are highly configurable by the user anyway. I guess mastodon doesn't fall under this category..

I haven't found a final version online.

I wrote an article / review of the drafts that were available a few months ago. You have the links to the sources.

https://outofthecomfortzone.frantzmiccoli.com/thoughts/2023/...

> containing the downsides of Silicon Valley’s latest export

through which political lense do i have to look in order to put it like this? there is decades of european (continuing the bloc thinking) research and braindrain (and wasted opportunity) going into this.

What are the regulations exactly?

If you can’t innovate, litigate.

"The algorithm" is going to be the next few decades' scapegoat. "It wasn't a human, it was a machine!"

"...coded by a human."

https://archive.is/FMB2W

Can't wait to read the deal.

But this should not be very hard to agree on, which pretty much showcases the fuckery going.

A fucking no brainer to turn down. Which it seems they not really did but perhaps did a little bit.

"The negotiators agreed to allow some live scanning of faces, but with safeguards and exemptions, Breton said. The deal would prohibit biometric scanning that categorizes people by sensitive characteristics, such as political or religious beliefs, sexual orientation or race. Officials said this was one of the most difficult and sensitive issues in the talks."

Seems impossible to store photos without recording religious affiliation. Some Jews, Sihks, Muslim women etc wear headwear that is a give away.

Storing a photo is not the same as "biometric scanning" of that photo to categorise it.

yay!

Good, fuck exploitative machine learning.

Great steps in the right direction. Now the EU only needs to ensure property is not stolen by ai companies. Essentially if you cant prove your model has been trained on content you have licensed then selling it should be illegal.

Guilty until proven innocent. Very thoughtful take.

It’s called burden of proof. And usually sits with those selling products.

In the US, the burden of proof is with the regulators, but there's higher liability than in the EU which is supposed to offset it. That usually applies to safety / fitness for purpose, though, not copyright: I don't think the arguments transfer from one to the other.

You're on a forum called Hacker News. Abolish all copyright.

Indeed. Copying Is Not Theft.

I could rephrase this way: Copying Is Theft if you made a lot of money with that. Copying Is Not Theft if you made a living of that.

There's no difference in that. Copying Is Not Theft.

But selling a copy is.

You're on a forum called Hacker News. Abolish all copyright except when it benefits me.

How about lets just get rid of all of it.

Copyright has been one of the primary driving forces behind human progress over the last couple hundreds of years. But sure, let's get rid of it... cause it's cool (or edgy, or whatever...)

People sharing or even stealing other people's intellectual property rights has been an even bigger driving force of human progress.

It's why places like San Francisco are the tech center of the world.

The reason being, that all of the tech employees will often get all this great knowledge and then leave and start their own company, or at a minimum transfer that knowledge to other companies.

The fact that California doesn't allow most non competes is an amazing example of how sharing knowledge has created amazing progress.

> People sharing or even stealing other people's intellectual property rights

Because they can make money doing that. There has to be balance, having no copyright protection at all obvious would be disastrous.

> this great knowledge and then leave and start their own company

That's great, it's not illegal.

> doesn't allow most non competes

Cool. Also tangential and has no relation to copyright.

> That's great, it's not illegal.

Indeed it is great that people can take knowledge and share it with others!

More stuff like that should be legal. Using other's knowledge should be expanded and should become even more legal by getting rid of laws that prevent information from being shared.

That's my whole point. It is an example where the sharing of knowledge and ideas is actually the large driver of innovation.

Stealing and sharing ideas promotes lots of innovation. Because workers take their knowledge and share it with new companies that they join.

> Also tangential and has no relation to copyright.

No it's not tangential.

It's completely related to the idea that sharing knowledge promotes innovation.

> Because they can make money doing that.

Indeed. People can indeed make money by sharing and stealing knowledge. That's my whole point!

It promotes innovation and allows people to even make money. Glad you agree.

> Indeed. People can indeed make money by sharing and stealing knowledge. That's my whole point! > It promotes innovation and allows people to even make money. Glad you agree.

Only because copyright laws and similar protections exist.

But I'm sorry, I can't really say that I agree with you because you're being so vague that I don't really understand what you are actually saying.

Not because it's edgy, but because it's the right and ethical thing to do. Free culture and free software, opponents of copyright, has driven much innovation over the last few decades.

> Not because it's edgy, but because it's the right and ethical thing to do

Also, utterly delusional. That is unless you don't see any problem living in a world with no movies, no tv shows, barely any books, a lot less music, video games etc. etc.

> Free culture and free software,

Opensource software has only been really successful in areas where it's a cost centre and not the end product. Which is why corporations are willing to fun its development. I'll let you figure out the implications of that yourself.

Sure, but abolishing all corpyright means that owned by corporations too. Let's see microsoft's source code. It won't be of quality, but as soon as that will be used for training ml i would agree to mine being used.

Why would they ever show it to anyone? Without copyright we'll go back to the middle ages with every craftsman/(modern equivalent) hiding their methods, techniques and technologies from everyone else the best they can (and consequently way less innovation). Also, we'd no longer have any movies, almost no books, music etc. bug I guess literally returning to the dark ages is what these silly "abolish all copyright" people want.

Right.. it's also called 'news.ycombinator' which is certainly not an organization which would ever support abolishing copyright. Besides it's also a terrible idea, not having copyright would stifle innovation for years.

It's probably one of the best things humanity ever came up with (good luck making money being a writer in the the pre copyright days...).

People like you must've thought science and art and writing and other shit didn't exist before currency. Like you need a monetary incentive to get a human to do anything.

And people like you must learn not to think (if you do actually think at all) in binary terms all the time (and maybe consider reading a history book or two).

Yes it existed, at a massively smaller scale.

> Like you need a monetary incentive to get a human to do anything.

So we agree? Or was that a question? Because the answer is yes. Without copyright it was almost impossible (and still largely is) for any creator to make a living without being independently wealthy/being funded by a third party.

If you're fine living in a world with almost no books (especially no fiction) and approximately zero non ultra low cost TV/movies (unless you count ads which would effectively become the only way of funding anything substantial..) etc. than sure maybe copyright is not really necessary.

Go EU, make "The Right to Read" real life!

The stupidest of the new rules:

1) AI Act includes bans on biometric systems that identify people based on "sensitive characteristics", such as sexual orientation or race (except for "law enforcement"). (LOL, so an AI isn't "allowed" to tell if a person is black, white or Asian?)

2) It also includes new transparency requirements for all general-purpose AI models and stronger rules for "very powerful" models. ("Yes, because with such 'great power' comes great responsibility).

>LOL, so an AI isn't "allowed" to tell if a person is black, white or Asian?

Yes. I don't see what you find so funny about this.

Race is one of the most important attributes used to describe a person of interest.

Sex, race, height, build, hair, last known clothing

And do you need an AI to guess your customer race or sexual preferences ? If is needed for the service (maybe you are a doctor)you just ask. If is illegal for you to ask and use someones race then it should not be legal for you to guess it and then use that guess in your algorithms.

If it is illegal to collect and use that information then you don't need to additionally regulate algorithms specifically...

And if it is legal to collect and use that information then it seems nonsensical to ban the use of algorithms specifically...

Nope. That's not how it works.

Even if it is legal to collect that information, it does not imply that it's legal (or desirable to begin with) to feed that information to a blackbox algorithm.

Hence I specifically wrote "use it"...

The GDPR use "process it", which is generic enough to cover AI.

> The GDPR use "process it", which is generic enough to cover AI.

Nope, it's not. Because, again, sending user data to a magic black box may be illegal. While processing it in other ways could be legal.

You are not replying to my comments. You are adding an additional constraint that wasn't in my initial comment.

My point is that if collection and processing is regulated then that covers AI. And thus there is no "sending user data to a magic black box may be illegal" but my whole comment was that this wasn't needed. AI is just a tool, it's just automated processing.

Btw, 'nope' passes off as rather dismissive and rude.

> You are adding an additional constraint that wasn't in my initial comment.

No. I'm showing that your conclusions don't logically follow from the premise.

> My point is that if collection and processing is regulated then that covers AI.

No. No it doesn't.

> but my whole comment was that this wasn't needed. AI is just a tool, it's just automated processing.

And that, too, isn't entirely true, precisely because AI is a tool, and a different one. We always adapt or pass new laws and regulations for new, significantly different, tools.

For example, just because you can process data doesn't necessarily mean that you can process it in any way by any tools available to you. "GDPR allows you process data" does not immediately imply "this data can be processed by AIs". There is a reason why I was "adding additional constraints". Because these constraints are why additional laws and/or regulations around AI are needed.

For example, one of the stipulations of the AI Act is [1]: "AI systems shall be developed and used in a way that allows appropriate traceability and explainability, while making humans aware that they communicate or interact with an AI system as well as duly informing users of the capabilities and limitations of that AI system and affected persons about their rights". The reason being stories like this: https://www.politico.eu/article/dutch-scandal-serves-as-a-wa...

That stipulation alone isn't covered by any existing laws. There are bits and pieces here and there about various processes in various contexts, but this act brings them all together in one coherent form applicable to AI systems (which are appropriately broadly defined in the act [2]).

Or, there's an entire Article 5 explicitly prohibiting the use of AI in sensitive and outright harmful areas. Including things like banning them for large-scale suveillance except in very narrow scopes. Note that, again, there's undoubtedly pieces of this scattered around various regulations, but this one makes them clear and make them specifically applicable to AIs.

That's why I keep repeating: no, the fact that someone allows to collect and process data does not immediately imply that this collection and processing can be done with AIs. Where applicable, this law references existing ones.

[1] Quoting from this excellent article https://softwarecrisis.dev/letters/the-truth-about-the-eu-ac...

[2] Annex I: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52...

>And if it is legal to collect and use that information then it seems nonsensical to ban the use of algorithms specifically...

And a HN entrepreneur will say that they do not collect the race, a neuron somewhere in the AI will turn on for a specific race but it is not their fault, the input data had no race the AI deduced it.

The laws regulate high risk stuff, it will prevent that HN entrepreneur to sell medical AIs , crime predicting AIs, car driving AIs etc just because they put a disclaimer that the AI might be wrong.

Collect it, store it, use it. QED.

That's why drafting of laws is so tricky and important. You need to find the right way to make a general principle apply without having to spell every single possible case, which is impossible.

But on those things I am convinced that there is no need to single out AI if the law is drafted so as to spell put general principles.

On the other hand, AI can bring major benefits and, while we should be careful, we should allow "tech entrepreneurs" to bring those benefits to society.

>On the other hand, AI can bring major benefits and, while we should be careful, we should allow "tech entrepreneurs" to bring those benefits to society.

When I read about this last time , the limits were n place only for risky domains. So AI for coding will not be affected.

Also the laws are not specific that the AI means neural networks, in my opinion from my experience tech is fulled with "fake it" people, with AI it is much easier for many more people to "fake it", so we can end up with this people selling crap to people and not taking responsibility for their actions.

I seen someone was working on a therapist AI, if he sells it as a service do you think that a simple ToS that says"the AI might be wrong, I take no responsibility" is enough? It is similar like Tesla AI, they also do not want to claim any responability and push the fault to the customer.

Race is a social construct. It's used the way you describe in some societies, but not in others, and even if it is, the races have different definitions.

It’s not entirely a social construct, for example there are very clear genetic factors - you could use a clustering algorithm on genetic data to reveal there are very clear groups, regardless of which social culture is doing this clustering, the clusters would still emerge, because math.

You’re right that the borders are not well defined though - it is a probably distribution - just like the weights in a neural network are also a probability distribution

There are no clear genetic factors at all. They don't exist, because there are no clear biological delineations between 'races'. It's a purely social construct.

You maybe could sort of define a group of people coming from the same locality and having the same history , and they would have maybe some degree of biological similarity as well, but there would be a high number of those groups and even then it would be quite useless and with number of exceptions.

Come on, that’s just today’s Lysenkoism. We can talk like adults here.

Prove me wrong. Give me a way to detect race from DNA sample with 100% accuracy.

Set the bar at impossible so you cant be proved wrong. No system anywhere is 100% accurate, and to demand that as a standard only shows your numerical illiteracy.

Why is it impossible? It's certainly possible to tell apart cat and dog DNA. If the race theory proponents argument hold water, the same must be possible for different races.

> Prove me wrong. Give me a way to detect race from DNA sample with 100% accuracy.

You implied that in order for you to be proved wrong, the test must have 100% accuracy. I was saying that 100% accuracy is an impossible standard.

No tests in the world are 100% accurate, let alone in biology, the fact that you asked for this shows you are numerically illiterate

Here is DNA clustering that shows race: https://4.bp.blogspot.com/-t9Va5ai9-A4/Xd0cmvRBL4I/AAAAAAAAu...

Here is a link to the Kahn Academy course on statistics: https://www.khanacademy.org/math/statistics-probability If you take this course, you will understand probability better, and then you wont commit the logical fallacy of "false dichotemy" where you try to reduce complexity to binary classes which is a view that is too simplistic to accuratly represent the distributions we are talking about

In that case races do not exist as meaningful biological categories. There is no proof of their existence.

The reason I mentioned cats and dogs is you can tell them apart by their DNA with 100% accuracy.

> Here is DNA clustering that shows race: https://4.bp.blogspot.com/-t9Va5ai9-A4/Xd0cmvRBL4I/AAAAAAAAu...

That image does not contain 'race', and it comes from a paper that does not even mention the word 'race'. You have provided no arguments that race exists. The paper writes:

> This observation ... will make it a challenge to trace the ancestry of African Americans to specific ethnic groups in Africa, unless considerably more markers are used.

The more markers you use, the more different groups you get. You can of course choose to 'cluster' them, but then some precision will be lost, and what exactly will be gained?

Ethnicity is not a social construct, and sometimes people use the word race as a synonym for ethnicity. There are clear genetic factors for ethnicity.

So you are saying you are able to clearly detect ethnicity of a DNA sample? Do you have anything to substantiate that?

That's literally what 23&me do, they send you a breakdown of what % of various ethnicities you are based on your DNA. E.g. 95% Northern European, 5% Ashkenazi Jew.

It's what they claim to do. The science is bunk. It's genetic astrology.

Yes and then they change it on each update. Do have a read of what their customers think about it, for example here: https://m.facebook.com/story.php?story_fbid=pfbid0cPeurGB4WX...

By definition, DNA is DNA.

Yes, you can see what someone's DNA is, if you look at their DNA.

But every innate characteristic encoded into our DNA is a social construct, and if you disagree you’re a nazi! Checkmate?

Ah yeah. I forgot that with a good frequency two black people give birth to a white baby or vice versa.

That's skin color, not race.

You don't see anything funny about a biometric identification system that is not allowed to consider factors that would be considered indicative of "race"?

Not really. It's reasonable to ban the automatic detection of stuff like this.

I honestly think it's stupid.

Person A won't be identified because of being Chinese, he will be identified because of being a second generation immigrant that has a Chinese diet, interest in STEM and has a Chinese name.

The reason there are racial stereotypes is because there are trends among different races and skin color is a very evident trait for humans to identify. AI will draw the trends from everything else instead, which will draw the same picture, but skipping one specific parameter because it feels icky for humans.

Assume an increased chance of someone being interested into the arts and liberal because of dyed hair? Sure! Assume an increased chance of knowing Spanish because of being Hispanic? Ew no that's prejudiced.

It's so very stupid.

A biometric id system which can only distinguish "clowns" from "non-clowns" and makes a lot of errors with regard to the images of famous billionaires.

That is funny.

Your example? Not funny at all.

As with anything the devil is in the details - The word “tell” there is doing a lot of work.

Does it mean “The AI should not have any bias towards any race”

Or does it mean “The AI should not be able to distinguish whatsoever between two races”

To me, the phrasing above “AI should not be able to tell” sounds like the latter, which is drastically wrong and I suspect difficult to prove.

I think most people would agree it should not have bias, but it’s important that in the wording of the law we don’t eliminate functional stuff that’s important - ie being able to determine race - for example there are certain medical conditions that affect different races at different rates, so the ability to distinguish is an important factor in analysis.

As user siminion314 wrote: "If [it] is needed for the service (...) you just ask"

I guess it's easier to require it not be able to distinguish at all than define what bias actually is and what the errors on that are.

The reason why it is funny is that almost by definition, anything that a biometric data detectors could detect would fall under the banned category.

A bio-metrics detector, detects biometrics! And the law basically bans bio metrics detectors from detecting biometrics.

It would be like if there was a law that said that restaurants can't sell food.

Or if hammer regulations prevented hammers from being used on nails.

Or that water could be used for anything but making things wet or drinking.

It is almost definitionally contradictory.

That is indeed funny.

It’s funny because 1) any person you interact with face to face can tell your race and 2) what is the AI supposed to do if you tell it your race?

I still don't see the humor here.

1) The overall point is to not let AI use racial bias as a factor in its decision-making.

2) Actually, there are a lot of people whose "race" is vague. My buddy has been asked if he's <random, different ethnicity/nationality> at least 5 times that I've noticed.

3) "Race" is a social construct anyways. In the US, in modern times, we delineate "races" in a specific manner, but these delineations have varied across history and cultures. See the one-drop rule (https://en.wikipedia.org/wiki/One-drop_rule).

You do not explain why these rules are stupid, much less "the stupidest".

Both rules, on first sight, make __a lot__ of sense.

“Transparency” is stupid for some people here because they cannot fathom that a technology significantly impacting people’s lives should be auditable.

Transparency of the inner workings isn't really relevant - getting the right results is (whatever that means in any particular situation). It diverts a lot of effort from solving the actual issues.

A transparent but poor decision is not really much better than an opaque poor decision.

I think that a world where that comment holds up to scrutiny would be very different to the one we've got.

There's no magic objective moralism box that you can put a decision into and have it weigh its consequences and a little green or red light turns on.

Good for whom? Under what assumptions was it made? In what context does it hold?

All of that is transparency.

That would be what regulation should address and not sidestep. Politics are there to make exactly those types of moral decisions. If you want certain behaviour from corporations using AI for something, it might be better to directly police the behaviour and not indirectly via peering inside tools.

Creating mountains of documentations on the model for a regulator will not do that.

Similarly, using the input/training data quality as vehicle to avoid, e.g., discrimination is weaker than directly targeting an outcome.

Like a ban?

That could be a reasonable decision for people to make. If observability comes with accountability then it might eventually end up as the same thing.

For some things, perhaps? Or some target accuracy/precision measures on others and fines for failing those. In the EU, individual harm can be quite cheap for corporations in some countries, that might also be an avenue.

Right, I can see your point now and my interpretation in my first reply wasn't very accurate at all. Sorry about that.

No problem. I probably wasn't quite clear enough given the many angles of the problem.

> A transparent but poor decision is not really much better than an opaque poor decision

But it is better, or rather more desirable, as it is easier to fix and testable against rules which might be very important in certain contexts.

Because it assumes the data only has negative use. Asian and black communities are underserved in many ways from employment, medicine to education.

Forbidding AI to identify race is forbidding AI tools to help serve those people better or even detect that they are being served poorly.

It assumes even that targeted advertising by race or gender orientation is only exploitative and not seeking to help fill key needs.

AI is only made by racists and religious jobs as a basic assumption seems like a very stupid assumption.

But as I understood it's already established that many of the big initial training sets were highly biased, so it's a legitimate concern.

It's illegal in most of the EU, maybe all of the EU, to register ones "race", even for the police.

What do you mean? I don't understand.

Reply on news.ycombinator.com