Apple Sued by Three YouTube Channels

[macduke]

Wait until those three YouTubers realize all AI companies steal from the internet

They’ve sued the others as well. And Ethan actually set precedent in court in past cases about fair use online. His legal team actually knows their crap and that’s what they’re trying to do here—set precedent so more creators of all kinds will have more protections online against these horrible AI companies that are taking our jobs while making everything more expensive and destroying the environment.

 

[scottishwildcat]

We're going to need some clarity in the law sooner or later. The content these creators contribute to the training model is a drop in the bucket. Their individual work is so trivial that omitting it entirely won't materially change the model. However, if we get to the point where models cannot be trained then we're going to end up in a situation where the astonishing capabilities unlocked with this technology will only be available to those with the most resources—and to those who live under more liberal copyright laws. American citizens may find themselves less and less able to leverage these groundbreaking technologies because we've been hobbled by our antiquated copyright laws.

There is nothing "astonishing" about having to babysit a stochastic token generator every day until it eventually spits out something vaguely plausible but still utterly non-deterministic. (And yes, that is now what some of my job entails.)

 

[contacos]

So, can they sue me if i watched their public videos learned something? Although, I guess in my case, I'll have gotten dumber watching YouTube.

No but if you printed out their work and decided to teach a class on it in the evening to bunch of people paying you for the course, they probably could

 

[WarmWinterHat]

So, public videos on a public site that anyone can view and use for reasearch yet ai is bad and the boogie man. All creators are influenced by others and some blatantly steal from others but ok.

Anybody. Again, AI is not a person and shouldn't be lumped with people.

 

[chrono1081]

I would love to see their proof on this.

EDIT: Disregard, I missed the part where it's in Apples research papers.

 

[contacos]

They are probably hoping for a settlement. I don't see how the courts would ever rule in their favor. How is this different from any publicly available information being used for profit? Search engines make billions off the results from every other website creator.

A) as a owner of a website, you can opt out by simply toggling "no-follow" in the CMS, no such thing with all these different AI providers. You probably don't even know they took it.

B) depending on your content, Google needs to pay them and there are publishing deals in place in countries such as Spain, France or Germany.

 

[staypuftforums]

It can be argued that Apple only exists today (in its current position) due to piracy and violating copyright.

The iPod saved Apple, and its success was due almost entirely to music piracy.

Without the iPod and then the iPhone, who knows where Apple would be.

 

[FatLouie]

What’s unethical about learning from a public video?

Essentially, and probably...that product is being "taken" from the content creator, and will be resold as information or content for profit by the owners of any, or all of the AI businesses. More or less.

I believe we can expect a ton of new laws, regulations, TOS, etc. for this AI adventure.

 

[spazzcat]

Essentially, and probably...that product is being "taken" from the content creator, and will be resold as information or content for profit by the owners of any, or all of the AI businesses. More or less.

I believe we can expect a ton of new laws, regulations, TOS, etc. for this AI adventure.

The issue I see is first you have to prove your content is truly original; otherwise, how do you prove XYZ AI model got its content from you and not the same sorce you used to create your content...

 

[WarmWinterHat]

The issue I see is first you have to prove your content is truly original; otherwise, how do you prove XYZ AI model got its content from you and not the same sorce you used to create your content...

That's why AI should only train on data it has explicit permission to use.

 

[lowendlinux]

Good keep suing, let's bury AI in court cases and make it unrealistically expensive to develop AI models

 

[jonparadise]

What’s unethical about learning from a public video?

Nothing, a person watching a video to learn something is perfectly normal.

But if AI scraping were a person, they would be doing the equivalent of watching the video, then downloading it, then chopping it up into chunks to use in their own video, or as a script, without permission or even attribution. For profit.

Just because AI does it on a vast scale, it doesn’t make it right.

A perfect example of how damaging this sort of thing is can easily be seen in real time by observing the number of websites going out of business since Google’s AI overview came in.

 

[groove-agent]

If successful, this could create an interesting precedent. The way I look at it, is people inherently learn though intimidation, from when they’re an infant. Most music from artists is technically a regurgitated mix of what was heard previously. However, if they can make a case where AI is different legally than an actual creative person, then this could be an issue for AI unless people agree to let SI learn from them.

Personally I think creating technology to fully replace the creative process in a human being we can do without. It will inevitably lead to a bad transition period for us.

 

[NufSaid]

Looking at Apple's "AI" - it doesn't appear they learned anything.

 

[firewood]

Essentially, and probably...that product is being "taken" from the content creator, and will be resold as information or content for profit by the owners of any, or all of the AI businesses. More or less.

I believe we can expect a ton of new laws, regulations, TOS, etc. for this AI adventure.

"Fair Use" legalities will change to allow this. The U.S. doesn't want to give away any advantage it has in AI technology to other technologically advanced countries that don't bother to recognize or prosecute copyright violations.

 

[spazzcat]

Nothing, a person watching a video to learn something is perfectly normal.

But if AI scraping were a person, they would be doing the equivalent of watching the video, then downloading it, then chopping it up into chunks to use in their own video, or as a script, without permission or even attribution. For profit.

Just because AI does it on a vast scale, it doesn’t make it right.

A perfect example of how damaging this sort of thing is can easily be seen in real time by observing the number of websites going out of business since Google’s AI overview came in.

Where you do you think these "creators" got their infomation from to make the video?

 

[WarmWinterHat]

Where you do you think these "creators" got their infomation from to make the video?

It doesn't matter, the creators are people and AI is not. They are not the same, and shouldn't be treated the same.

 

[The Phazer]

I assume they are suing Google too. Gemini definitely uses YouTube videos.

The terms and conditions of YouTube actively granted Google permission, so no case there has a chance.

Ultimately the situation in the US is that Congress should legislate here but clearly aren't going to, so the copyright status of AI training is really only going to be solved by a few Supreme Court decisions on if it's an infringing act or not. This case is as good as any for bringing that about.

*However*, what might present a further wrinkle is that a few courts recently have found that circumventing YouTube's "rolling cypher" is a violation of DMCA 1201, and if that held in this case (and the no doubt higher court appeal) then anyone training on YouTube content that wasn't Google would be absolutely screwed legally, including Apple, even just looking at existing settled law.

 

[coolfactor]

Do they have hard evidence? They stated this...

Apple's research papers indicate that some of the YouTube videos uploaded by the plaintiffs were used to train its AI models, the complaint alleged.

Which papers specifically? Are these included in the suit with highlighted extracts?

 

[spazzcat]

It doesn't matter, the creators are people and AI is not. They are not the same, and shouldn't be treated the same.

Not sure what you are trying to prove, but your right AI is not a person, so how would copyright apply to it.

 

[WarmWinterHat]

Not sure what you are trying to prove, but your right AI is not a person, so how would copyright apply to it.

You apply it to the company, and they legally a 'person'.

 

[pigeonguy]

Another fail by Ethan “pedo troll”.

Hard agree. The second I saw him as a plaintiff I immediately dismissed his claims. The guy just isn’t reliable and shouldn’t be trusted.

Not sure who the rest are, but he can suck eggs lol.

I wonder what evidence they have of this.

 

[visualseed]

I'm guessing the YouTube ToS has language absolving Google and their partners of liabilities for use of their content.

 

[visualseed]

The terms and conditions of YouTube actively granted Google permission, so no case there has a chance.

Ultimately the situation in the US is that Congress should legislate here but clearly aren't going to, so the copyright status of AI training is really only going to be solved by a few Supreme Court decisions on if it's an infringing act or not. This case is as good as any for bringing that about.

*However*, what might present a further wrinkle is that a few courts recently have found that circumventing YouTube's "rolling cypher" is a violation of DMCA 1201, and if that held in this case (and the no doubt higher court appeal) then anyone training on YouTube content that wasn't Google would be absolutely screwed legally, including Apple, even just looking at existing settled law.

Or a Google partner. Apple could very well ask for indemnification in any current or future agreement with Google.

 

[DocMultimedia]

Who? Whatever. You post stuff and you should just assume it's getting seen by everything from dumb humans to dumb AI. Siri, can I watch this video? I'm sorry, I don't know how to do that.