Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

Reality

13 (68.4%)

Universe

3 (15.8%)

The World

3 (15.8%)

Total Members Voted: 19

[EuchreJack]

Most AI generation of music and images violates US Copyright Law. I'm mildly curious how that is going to be resolved.
It may be that Sony will eventually own several AI generation programs after suing their current owners out of existence.

[Strongpoint]

Most AI generation of music and images violates US Copyright Law. I'm mildly curious how that is going to be resolved.

Hm, which part? It is a very questionable take. I wouldn't claim this until we get enough relevant court decisions or new legislations.

I expect Sony and Disney and their friends will try to demand the licensing of training data but I can't see how it can be done without a total destruction of the fair use doctrine.

[EuchreJack]

Fair use doctrine is limited to non-commercial uses. Since most AI is arguably for-profit, or could be used for-profit, it's a minefield.

[Strongpoint]

Fair use doctrine is limited to non-commercial uses. Since most AI is arguably for-profit, or could be used for-profit, it's a minefield.

Who told you that? You are wrong.

It is way easier to get Fair Use for non-commercial uses but it is absolutely possible to get Fair Use for commercial uses. Parodies wouldn't exist otherwise. Many types of reviews wouldn't exist either. YouTube would be a sad place if commercial Fair Use didn't exist.

[MaxTheFox]

Honestly, we will be getting more and more grey areas like this as time goes on and new tech appears. Copyright law, in its current form, is outdated and barely functional.

[EuchreJack]

YouTube also blatantly violates US Copyright Law...

But yes, reviews and parody are valid commercial applications of the Fair Use doctrine. Neither applies to AI generated words, since they don't explicitly reference to original work.

AI generation is outright copying that pretends it is not.
None of the original artists, not the companies that bought their souls, are receiving any credit or income for the images/sounds that were imputed into the machines. And the machines can no currently work without those inputs.

[MaxTheFox]

AI models don't actually contain the inputted works, is the thing that causes it to be a grey area. I can't see myself personally caring about my writing being used in AI training tbh.

[EuchreJack]

AI models don't actually contain the inputted works, is the thing that causes it to be a grey area. I can't see myself personally caring about my writing being used in AI training tbh.

Good point. I imagine that sort of argument should keep the AI run by the richer folks alive.

[lemon10]

Agreed, it is not in fact outright copying, that simply isn't how AI works. AI training is logistically very similar to human learning in how it works, something which is very much not illegal.
Because of this it isn't really a minefield, the laws of copyright don't actually make it illegal. It really *feels* like it should be, but the law doesn't work based on how things feel.


That isn't to say there aren't serious legal issues around AI. The whole "lets just strip-mine the whole internet for stuff we don't own, much of which is not publicly available and now says explicitly says you can't use for AI" is substantially more legally dubious. The first part seems to probably be fine as long as it was posted publicly (again, legally), but the second half is still up in the air.
---
There are decent odds that the copyright will catch up fairly soon simply though cause big corps don't like people taking their stuff without paying, but given all the AI money on the other side its not a sure thing.

https://twitter.com/front_ukrainian/status/1781968599243989420

Killerbots are coming

Note that Russia already tried some similar AI targeting assist for their weapons (albeit just for tanks instead of people). It just sucked so they removed it lol.

spoiler=Two relevant xkcds-- not just to AI but to the general mindset here

...I already had a few of them in mind
https://www.xkcd.com/1007/ - shows where a logistic curve might be more apt than a logarithmic one, sometimes
https://www.xkcd.com/1281/ - sometimes not actually necessarily wrong (nor the title text)
https://www.xkcd.com/2892/ - a problem with all such extrapolations
https://www.xkcd.com/2914/ - let's call this, in AI context, the "not-uncanny ridge"


(Also I had in mind something about both Black Swans and Grey Rhinos, that might be needed to fulfil the promises, but they're respectively the unknown unknowns and unknown knowns...)

You missed the best comic on exponential growth though.
https://www.smbc-comics.com/comic/2011-10-14

Spoiler: Two relevant xkcds-- not just to AI but to the general mindset here (click to show/hide)

"By the third trimester, there will be hundreds of babies inside you."


"2031: Google defends the swiveling roof-mounted scanning electron microscopes on its Street View cars, saying they 'don't reveal anything that couldn't be seen by any pedestrian scanning your house with an electron microscope.'"

Yes yes, I know you don't believe that exponential growth is real.

[Strongpoint]

YouTube also blatantly violates US Copyright Law...

But yes, reviews and parody are valid commercial applications of the Fair Use doctrine. Neither applies to AI generated words, since they don't explicitly reference to original work.

AI generation is outright copying that pretends it is not.
None of the original artists, not the companies that bought their souls, are receiving any credit or income for the images/sounds that were imputed into the machines. And the machines can no currently work without those inputs.

Citation needed.

Also, the New York Post is suing Microsoft right now so I wouldn't be so sure about text generation being any different than music or image generation. In fact, no publically available model in existence cited a full copy of a copyrighted work (part of it) in the output. Language models do this all the freaking time.

If YouTube is breaking the copyright law why Google is not sued into bankruptcy?

US list of allowed uses is not exhaustive, you can't say "yep parody, review and education are allowed but not the thing you suggest". All 4 factors of fair use need to be evaluated before you can claim that.

Do you know that under the US copyright law I can take a bunch of images, glue them together in a collage, draw some stuff over it, and exhibit the resulting creative work for money? And it all may be the Fair Use depending on the evaluation in the court. Which depends on how creative it is, the extent of the stuff I added, what % of copyrighted works I used, and many other factors.



For image generation models. I'd say that both 1st and 3rd heavily weigh in favor of Fair Use evaluation, while 2nd and 4th weigh in other direction.

I think that heavily transformative use (Images and software that draws images are very different things) and the fact that copyrighted data is incredibly diluted (except in cases of blatant overfitting, even then it is diluted) We are talking about a ridiculous amount of images + tags compressed into mere gigabytes. It would be a very, very lossy compression even if it contained no original data but obviously, it does contain other data except heavily converted and compressed images.

It is nowhere close to being a clear example of Fair Use but there are good chances that it is. And, most definitely, it is not a clear example of not Fair Use.

[EuchreJack]

YouTube also blatantly violates US Copyright Law...

But yes, reviews and parody are valid commercial applications of the Fair Use doctrine. Neither applies to AI generated words, since they don't explicitly reference to original work.

AI generation is outright copying that pretends it is not.
None of the original artists, not the companies that bought their souls, are receiving any credit or income for the images/sounds that were imputed into the machines. And the machines can no currently work without those inputs.

Citation needed.

https://guides.lib.usf.edu/c.php?g=1315087&p=9690822#:~:text=Generative%20AI%20tools%20can%20be,may%20need%20to%20be%20obtained.
https://hbr.org/2023/04/generative-ai-has-an-intellectual-property-problem
https://www.axios.com/2024/01/02/copyright-law-violation-artificial-intelligence-courts
https://theconversation.com/generative-ai-could-leave-users-holding-the-bag-for-copyright-violations-225760
https://crsreports.congress.gov/product/pdf/LSB/LSB10922
https://www.theverge.com/23444685/generative-ai-copyright-infringement-legal-fair-use-training-data

[MaxTheFox]

Yes yes, I know you don't believe that exponential growth is real.

Correction: it's real, but by its nature it never lasts long in any real environment. Permanent or very long exponential growth exists only in mathematics and in subpar sci-fi*.

*written by people who know more far more math than they do history and sociology

[Strongpoint]

YouTube also blatantly violates US Copyright Law...

But yes, reviews and parody are valid commercial applications of the Fair Use doctrine. Neither applies to AI generated words, since they don't explicitly reference to original work.

AI generation is outright copying that pretends it is not.
None of the original artists, not the companies that bought their souls, are receiving any credit or income for the images/sounds that were imputed into the machines. And the machines can no currently work without those inputs.

Citation needed.

https://guides.lib.usf.edu/c.php?g=1315087&p=9690822#:~:text=Generative%20AI%20tools%20can%20be,may%20need%20to%20be%20obtained.
https://hbr.org/2023/04/generative-ai-has-an-intellectual-property-problem
https://www.axios.com/2024/01/02/copyright-law-violation-artificial-intelligence-courts
https://theconversation.com/generative-ai-could-leave-users-holding-the-bag-for-copyright-violations-225760
https://crsreports.congress.gov/product/pdf/LSB/LSB10922
https://www.theverge.com/23444685/generative-ai-copyright-infringement-legal-fair-use-training-data

Thank you. The very first link proves my point.

As of April 2024 several law suits have been brought against AI image and text generation platforms that have used visual and text content created or owned by others as training material.  These law suits claim that the use of artists’ or writers' content, without permissions, to train generative AI is an infringement of copyright.

While these cases are ongoing, we have no definitive answer on whether the training of AI models is considered an infringement of copyright.  However, several experts have pointed to previous fair use cases to justify a fair use argument for the use of various training data for AI image generation tools.

It is exactly what I am saying. Experts say it is very possible that training of AI models is fair use. While you claim that it is definitely against the US copyright law and fair use has nothing to do with it because... reasons.

What will the courts decide? No one knows, it is a battle of money. But if they will decide that those are not fair use, the fair use doctrine will receive a huge hit.

[King Zultan]

YouTube also blatantly violates US Copyright Law...

I have been given this isn't the case as they often go through and take stuff down that is against copyright.


It's also the reason I've been downloading entire channels, as several videos I liked were got for copyright stuff.

[eerr]

Also, who wins in the copyright debate may depend on what the big lobbies think. This could go on for a long time if you have big companies on both sides.

Remember, historically, businesses put forth copyright, to make sure they can make money off what they create.
Also note there is a decent number of non-ai businesses that see some little guy's art and just steal it. So this type of thing being theft sort of has precedent.

The little artists need to pop up and give the courts an opinion that isn't 'this is going to take my job'.

What would be really funny, if it the final decision is that only one company in the data daisy chain has to pay all the artists they stole from.
But only once, covering all companies' use of the dataset.
So little Timmy gets five dollars per art piece and then 500 companies proceed to use his data for ten thousand years.