LinkedIn does not use European users' data for training its AI
techradar.com
techradar.com
There are some consumer protections that I really do wish we imported into the U.S., especially food safety and chemical usage. Too much regulatory capture for that, though.
If they didn't want these resources to be exploited, they shouldn't make them publicly accessible!
Automatically doing something is a whole other quality from a person doing it. Police watching a protest is fine, police filming it or documenting all participants via face recognition is forbidden (at least here).
Whatever their terms say, they're storing your stuff, they are serving your stuff, and they've reserved the right to extract value from your stuff in perpetuity, according to your agreements when you signed up, when you posted, and when they updated. Doesn't matter about your privacy settings, because it all happened on LI.
I mean, you can delete the stuff you posted if you don't want future AIs trained? Delete your whole account if you didn't like LinkedIn messing with it?
(I could likewise say this about MS Windows, Apples or anything: if you don't want someone to have your stuff, give it to someone else, or don't give at all?)
But in the end, you voluntarily gave it to them, because it was free, but you are the product, and not the artist.
Rules change for monopolies and oligopolies, and for a good reason IMHO. LinkedIn belongs to Microsoft, so does GitHub. Don’t forget Windows, Azure, Office, Visual Studio and a long list of other products. They want to take your data from all possible sources and if you just point to the TOS alone this would be totally valid. But we have to look at the bigger picture and already do so in other areas, for example GDPR.
The point of making legislation is to have things to enforce at times like buy outs to say things there is no reasonable way to enforce our expectation that our 2FA numbers are not abused by this new buyer so the buyout can not continue.
Maybe you don't need a job social network, but presuming you do you have no way of knowing what org will own the physical media of the one you pick today unless it is in a country with competence, I.e. not the US.
I've said this before, but it's sad how quickly this community swapped from being champions of the free exchange and use of information for the betterment of humanity to gleefully stomping on an incredible and beautiful new technology because someone else might make money off it. Reminds me of the Judgement of Solomon [1] (people would rather kill the whole technology and all the incredible things that may come with it then miss out on "my cut!", "my cut!" even if it's a single LinkedIn post in a corpus of billions)
well, corporate use of technology for mass surveillance is equally incompatible with the idea of a democratic society, which is why the EU imposes limits on what LInkedIn can do with your data, and thank God for it.
Free exchange of information is being able to access a textbook at the library, not a corporate behemoth vacuuming people's personal information so they can sell them to the highest bidder to put you under surveillance the next time you attempt to switch a job or send you more ads. Betterment of humanity? Blink twice if a LinkedIn PR person is holding you at gunpoint
I was referring to Generative AI in general, this use-case is quite boring.
> a corporate behemoth vacuuming people's personal information so they can sell them to the highest bidder
Do people not use LinkedIn to explicitly signal to the world that they're looking for a job? Why does it matter how that information is being delivered? If you don't want the world to know you're looking for a job simply don't update LinkedIn?
> corporate use of technology for mass surveillance is equally incompatible with the idea of a democratic society
The argument against government power/surveillance is that they have a monopoly on it and may use their power to hurt people. It is good to legally protect sensitive information like health data from advertisers, but in this case you can, again, simply not use LinkedIn. What difference does it make if the info is collected by a company looking for new hires, a third party analytics company working on behalf of them, or LinkedIn itself working on behalf of them? It's not private data.
This applies for pretty much everything in our daily life like banking, shopping etc... "just don't interact" is such a useless nothing-burger that side-steps the problem entirely. You can "solve" all societal problems by becoming a hermit, moving to the woods and living off the land... but that is not a functional or reasonable thing to do for 99.99% of people. Its baffling especially when the reasonable solution is simply having a bare minimum standards of protection across the board, which many countries already implement to great effect.
Wait what, this isn't even close to being true
> government power ... is that they have a monopoly on it
This is only true by definition. McDonalds also has a monopoly on serving McDonalds food.
> use their power to hurt people
This is universally true of anyone with power. The difference between the government and other powerful organizations is that the government has a universal feedback mechanism via voting. In a democracy everyone gets a say in the laws that affect them.
I genuinely don't see what the problem is.
I rarely post updates on LinkedIn. When I do, they're updates that are intended to be broadcast to the public. If some execs at LinkedIn are smart enough to find a way to profit off the back of that, why should I be upset about that? Why are you upset about it?
This sort of language is unproductive and causes further division. It's just buzzwords hoping to evoke a certain emotion in the reader. I'm being genuine when I say the following:
> corporate behemoth
Why is this bad? (Steelman: it isn't bad by itself, but it makes the other things worse.)
> vacuuming
Evokes a certain image, but it is not true. It is copied.
> personal information
True, but note that it is public in this case. Why is this bad? What's so special about personal information (I'm interpreting this as PII specifically)? (Steelman: it can be used to track you; see "under surveillance the next time you attempt to switch a job")
> sell them to the highest bidder
Why is this bad? (Steelman: it feels unfair, to have contributed something without compensation; Counterpoint: access to LinkedIn itself is your compensation, such fairness is a subjective feeling, compensation is for example not needed if you reuse GPL code in a commercial context)
> under surveillance the next time you attempt to switch a job
Is this true? If this is true, I can see why LinkedIn AI would be bad, but only for this specifically. If this becomes false, then it's no longer bad. I doubt it is true, however, since this is not generative AI, which they're likely focusing on.
> send you more ads
This has nothing to do with generative AI, which I assume we're talking about.
> Betterment of humanity? Blink twice if a LinkedIn PR person is holding you at gunpoint
Not a useful statement.
This statement can be interpreted as:
1. Some AI is bad (e.g. "some AI take jobs"), therefore LinkedIn's AI is bad. This is an invalid argument, so we can ignore this.
2. All AI is bad, therefore LinkedIn's AI is bad. Valid, but probably unsound, because intuitively this would mean saying "AI used to fight cancer is bad". So we can ignore this too.
3. LinkedIn's AI is bad by itself. Why? What is LinkedIn using it for? Is it just a chatbot? There's no risk of obsolete careers from that. Is it a privacy issue? It is public data. Is it because it makes money? What's wrong with that? Is it simply a fear of the new? I think so, but that's just my uncharitable opinion.
Please elaborate.
Well, for context, when the US wanted to create a mass surveillance apparatus in the 1990s, they funded the tech sector. Now the tech sector both does the surveillance and fuels the economy. The surveillance info is fed back to the government in various ways and is increasingly used for things like military targeting.
Generally speaking, any information you give to a private company you should assume will be used by your government. Either directly via subpoena/tapped data lines/etc, or indirectly via AI services provided by these companies to the government.
You're assuming this community is a monoculture. I'm sure you'll see people on various sides, from ardent free exchange advocates to die hard copyright supporters.
To me, one could call this technology incredible and beautiful in the same way that a detached observer might use those terms to describe a hurricane bearing down on my hometown. In the Kantian sense it is more "sublime"[0] rather than "beautiful", and yet the sublime is experienced as such only if it can be viewed from a position of safety; but there appears to be no safe place from AI.
[0]: https://1000wordphilosophy.com/2023/11/11/kants-theory-of-th...
“No secret” from a human reading it is something categorically different from a machine ingesting it en masse and retaining it forever.
In other places, LinkedIn silently opted the users into AI training. From few days ago:
https://news.ycombinator.com/item?id=41582951 - LinkedIn scraped user data for training before updating its terms of service
https://news.ycombinator.com/item?id=41584929 - LinkedIn silently opts users into generative AI data scraping by default
The EU is structurally difficult to effectively, ah, lobby. This, in my view, is the big differentiator; if you're a company who wants to avoid inconvenient regulations, you don't just have to compromise one political party. Unless things are very, very close, you'd have to get to, at least, tens of national parties, plus a decent bloc of MEPs, to really screw things up.
Also the EU doesn't have almost any any successful consumer tech companies (and the entire sector itself is small and has little real influence).
On the brightside this leaves a lot of leeway to pass laws that protect consumers, regulate and limit monopolies etc. There won't be any domestic economical/political impact if MS, Google, Apple make a bit less money from ads, services and other stuff. So politicians have no reason to care about them.
The few billions in occasional fines might not be a big deal on the whole but it's still extra revenue (and the EU budget isn't really that huge). e.g. the ~1.5 billion Google antitrust fine is almost 1% of EU's yearly budget. GDPR also provides a constant trickle of revenue (https://www.enforcementtracker.com/) of course it's still pennies compared to the economic benefits of having a few successful tech megacorpos in Europe would have but I guess better than nothing.
It's always shocking when you find that some brand-name product tastes different in the EU because an ingredient used elsewhere is not considered apt for human consumption here.
After living a few years in The Netherlands, a lot of Argentinian candy has a faint taste of something kinda chemical. A few others who've moved here expressed perceiving the same flavour.
My guess so far is that living in Argentina we're continuously exposed to whatever this ingredient and become desensitised to it. When we live abroad for enough years, we become sensitive to it again.
It also means that you can't keep you margins really low as your risk remains high as it's turnover based.
https://www.linkedin.com/mypreferences/m/settings/data-for-a...
Why drop the case? A large amount of data was still collected, was it not? I'm sure X is happy to agree to not collecting data in the future, seeing as how they've already collected, what, 18 years of data already?
I'm of two minds.
If tomorrow a user facing feature needed to be radically different between US and EU users, I think they'd also delay EU rollout until it reaches an actionable priority, so it's kinda par for the course.
You could use it to indicate the fit between employees and companies;
You could use it to detect lies and exaggerations in CVs;
You could use it to estimate when employees are likely to be considering seeking new opportunities.
knowing LI doesn't have any NSFW content, is full of marketing content (both corporate marketing and self boasting individuals), tends to prefer positive signals (all projects succeed, etc.), is mostly done in English, and so on...
... There is clearly a market for text generation in this language bubble. Think of all the internal and external communication in your $BigCorp. That has immediate use not only in marketing, but also HR, recruitment, company policing, etc. Your next company town hall can be prepared and led (!) by this thing.
Make of that what you will
Someone who created their account from EU IP-address, someone who says they live in EU, someone who says the work from a company (division) based in EU?
Given the UK’s privacy and data protection environment is still largely the same as the EU’s, I wonder whether this was an oversight.
Who will ever know if they do or not...
The average webdev in many countries of Europe earns pennies and cost of living being still abysmal. Meanwhile one can only be breathing in the USA with a CS degree and easily make quite a lot, more than enough to get by.
I don't really mind that I earn less here.
But consider the other side of this coin : one of the biggest risks identified for AI is bias in training sets and there are actual demands that companies explicitly make their training sets as inclusive as possible to incorporate all cultures, genders, etc etc.
So if Europeans find out they are being excluded from job opportunities later on because employers are using AI tools within LinkedIn to process candidates and it simply doesn't understand the background of European candidates - will they be upset? Will they be demanding LinkedIn be fined for not including Europeans in the training set?
I am very curious how all this will play out long term as these competing tensions get worked through.
The whole idea of "you cannot sue me for my discriminatory practices, because I colluded with AI company X to hide the process" is a net negative. It does not matter which side of the coin is up, when I have to pay for it.