Hacker News Reader

Comment by pizzly | Hacker News Reader

As software, creative works, science, etc become more and more contributed via AI does that mean all future works will not be copyrighted or patented? Under our current copyright culture and laws obviously not but that does open the question to how much does a human have to contribute and what evidence is required to show that a human contributed enough so that it can be copyrighted and patented. Some time in the future (may be many years) science may become to complicated for humans to understand fully with AI systems researching themselves. Current patent laws in many places including US says inventions created entirely by artificial intelligence cannot be patented. This future may be coming. What will the new copyright and patent laws look like in the future? Do we need copyrights and patents anymore?

View on HN

It means as long as you can find a few employees who testify that they made "creative input" during the process and nobody can disprove that, you're fine.

It's like things that are already in public domain. Even if you make a coloring book out of paintings in public domain, it doesn't necessarily mean others can just print your book as-is.

Like being an executive producer

Perhaps but what about individual works done not via a company. Say you are a author and someone falsely detects a AI watermark in your book. Is the testimony you provide enough to override that falsely detected AI watermark? My feeling is that there will be an industry that arises that proves you contributed to the works you want copyrighted.

(I think) the case depends on what AI watermarks are. As far as I know, things like SynthID has practically zero false positives.

My understanding is that you get back a percentage from SynthID regarding how confident it thinks the text is generated by AI or not. The question is what percentage would be considered undeniable that it is AI generated or require more than just your testimony to prove that you actually did the work yourself. https://www.nature.com/articles/s41586-024-08025-4 (Google Deepmind) was using a "Watermark detectability is measured using the true-positive rate (TPR) when the false-positive rate (FPR) is set to 1%" for instance in their research paper. This 1% was set arbitrary just for the research paper but what percentage should it be set to?

I think you need to provide independent sources for the "SynthID has practically zero false positives". Except a paper written by Google authors about SynthID for images I have not found any study about zero false positives.

You can blame the governments for all sort of inefficiencies but any operation that impacts taxes runs a tight ship in most developed nations, this include things like copyright law, census, and anything statistics.

If companies get slope with creative output to the point that "a few employees" can reproduce it in shadow markets, don't expect to get copyright protection without giving governments revenue and speech-control.

> It means as long as you can find a few employees who testify that they made "creative input" during the process and nobody can disprove that, you're fine.

Have you been involved in copyright or patent litigation?

Because it's not that easy.

It's usually easier than that, as one by default hold the copyright of their works. Typically a company doesn't need to provide the whole working process and files to the court to claim copyright over their published products / to accuse pirates for violation.

If you're on the defendant side of a copyright violation case, it's extremely hard to use "well the original author didn't really make it...* as a defense. (Patent cases are often defended with this argument though, as a patent grants far boarder protection than copyright and can be rejected on prior art. But still it's very different from "AI made this actually.")

> It's usually easier than that, as one by default hold the copyright of their works.

That's not what we're talking about.

This is a question of copyright eligibility in the first place. An excellent recent(ish) example is Oracle v Google wherein the former claimed copyright over the Java APIs. That particular litigation went on for years and ultimately went to the Supreme Court (though it ultimately ended on a fair use argument and the issue of API copyright eligibility was never settled).

We can probably assume similar high profile cases in the future that turn on quantifying the level of human contribution in code produced through the use of AI.

As for patents, it's literally that easy. Attacking authorship of a patent is an excellent way to invalidate it. It's why any company with a serious patent program requires staff to keep supporting notes for their work and confirm authorship in writing as part of the filing process.

"Content that is entirely generated by artificial intelligence is not protected by copyright."

If that is faithful reading of the law, that makes sense. I know a number of people who use AI, but none of them (that are making anything actually useful) have the output "entirely generated" (aside from some POC tests that never see the light of day).

I have a hard time believing anything of value, anything worth copyrighting, could be entirely generated by AI.

> I have a hard time believing anything of value, anything worth copyrighting, could be entirely generated by AI.

Perhaps you're not stretching your imagination enough. What if an expert novelist used an AI like it were a fancy auto-completing dictation machine to write the next great American novel? AI may have "entirely generated" all the text, but what if they micromanaged the shit out it?

I can imagine the difference between someone who fires off a lazy 5 minute prompt, and someone who labors for months and months to get exactly the results they want.

It’s almost like laboring for months and months would have draft revisions, dozens of changes, and many other interactions they could prove as theirs defeating your entire argument.

The current trend (since months now) is to have everything done through agentic loops. Meaning humans are only here to provide the initial prompt and do a few back and forth during the implementation.

Most of the people I see doing that ship nothing. They just seemingly loop for the sake of looping.

However, I just saw a product yesterday that was released and being built as you described. Interesting times.

This is less relevant for logos & creative works, but things that enable companies to do production will go back to being closely guarded and sharded secrets, which is what the patent system was trying to resolve (though we can certainly argue the success of it).

If companies fail to protect their investments in generating IP, they will stop investing in generating it.

And unless IP generation costs (all in, including the humans telling them what to generate) fall close to zero, it will be bad for the world if companies cannot recoup investments in generating new IP.

We would expect this to hit those industries relying on IP protections the most, e.g. pharma.

Pretty much my thought: <https://news.ycombinator.com/item?id=49388777>.

> As software, creative works, science, etc become more and more contributed via AI does that mean all future works will not be copyrighted or patented?

No, just the stuff from people who think whatever they prompted from an AI is a contribution to creative works, science, etc.

> Content that is entirely generated [...]

Take note of the qualifier entirely. If you're working with an agent steering it to produce the results you want, it would be an entirely different story.

Where is this qualification in the legal language?

Where is the legal language?

Exactly, I don't know how the person I was replying to was making such confident statements about the situation.

I think that’s simple. The submitted title is more absolute than the thing it leads to. The article is talking about content that’s exclusively generated by AI. That’s very different than work that contains, or is derived from, AI output, or anything along those lines.

Wheres the limit. Would a prompt "Create a system that does x" and no other feedback count? Or do need to do 10 prompts to steer your agent to the results you want. What about just 1 additional prompt. Somewhere there will be a line.

Legal systems are fuzzy. The limit will be what courtrooms decide. There will never be a ">75% of content" law, everything will be left to interpretation, laws are not source code.

Is working with an agent different than constantly editing an image with fine tuning prompts? Both can be entirely generated.

a bit further down ...

> Neither mere prompting nor the selection between several AI suggestions is sufficient as a human creative contribution.

which reads like you actually have to contribute by modifying the output from the system, i.e. prompts/inputs to the system do not count as a human contribution.

If you modify the output, you don’t magically get copyright on the whole output, only the section you modified

if you modify enough of the output as to make it unrecognizable as the machine output -- that's a new work created by a human. i.e. you don't have to change every line.

same thing applies to sampling in music. if i change enough of a sample as to make it unrecognizable, then i don't need to worry about registering/licensing the sample.

Yeah, but then you end up more or let rewriting the LLM output, which defeats the purpose of using agents, no?

for anything you want to be able to copyright, yes.

otherwise you can still prompt your agent things, but any output falls under public domain.

since this hype wave started some people have been saying don’t use these tools in an environment where you’re dealing with proprietary IP. did anyone listen? no. because shiny.

More importantly, take note of:

> Munich Local Court has held that

Local Court in Germany is the equivalent of the Magistrates Court in the UK or Australia.

Definitely not the arbiter of ultimate truth.

Why would copyright and patents exist when the US blatantly stole the entire world's knowledge to train their AI models (and China distillled from it)

Knowledge cannot be "stolen"

You wrote your house's security code on a Post-It note. Somebody came in as a houseparty guest and located the note, and discretely copied it onto their hand to get access to the house during your next vacation.

Which verb would describe their knowledge of this security code?

"eavesdropped"?

"spied"?

Whatever it is, slapping "steal" on it because what was done is wrong, and stealing is wrong, is the same sloppy logic that the copyright lobby used in the 80s/90s.

If 'steal' doesn't fit, then neither do 'eavesdrop' or 'spied'.

learned

After a bit of thinking, I agree that "learned" is the most appropriate. I stand corrected.

The crime there is using that knowledge illegally to break in.

Obtained, read, viewed, learned, discovered.

The knowledge wasn't obtained illegally or illicitly, immorally perhaps.

Illegal entry/trespass is a crime regardless of the method used however.

In your analogy copying the code (online content) isn't necessarily illegal, however using that knowledge without compensation, permission or attribution may be a crime (depending on copyright law, judges rulings etc).

This point makes sense to me. I agree with "learned" and "discovered".

https://blog.curiousquail.com/im-upset-again-about-a-co-crea...

What about using torrents to acquire that content?

Torrents are legal?

Seeding is often considered the actual point of infringement, aka sharing copyrighted material without the right licence.

I should note I suppose that I don't agree with companies using artists or developers hard work for their own financial gain without compensation, whatever form that may take.

The problem as usual is the availability of legal loopholes when you have a team of lawyers and near infinite money.

It really depends the country, downloading a Hollywood movie via torrent is itself illegal in a lot of countries. The only exception I’m aware of is Switzerland

You're right I was thinking of civil vs criminal penalties.

Personal downloading is generally only a civil violation, requiring the rights holder to sue you for reparations.

What about art?

I think that you might be taking "knowledge" too literally. As I read it, the parent's post was using "knowledge" poetically for the various works that AI companies stole.

The more charitable interpretation would be "knowledge embodied in copyrighted works collected and duplicated without grant of rights".

<https://news.ycombinator.com/item?id=48821172>

Reply on news.ycombinator.com