Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

OpenAI may not use lyrics without license, German court rules

reuters.com

Post reply on HN
Print view View original

201–210 of 256 posts

Re: OpenAI may not use lyrics without license, German court rules

#201

Post by jeroenhd »

techblueberry wrote: post #79
petesergeant wrote: post #11

I am curious what happens if they call their bluff on this and cut off ChatGPT in Germany. Not that I think OpenAI is doing the right thing, just, I don’t think a country’s government can justify no commercial LLMs to its populace.

German student performance will plateau, while all other countries slowly decline.

AI is actively harming kids' abilities while inflating their grades when they make AI do their homework.

German student performance may plateau, but when student performance in other countries falls, that still leaves them in a better place.

Re: OpenAI may not use lyrics without license, German court rules

#202

Post by ghssds »

rpdillon wrote: post #148
EdgeExplorer wrote: post #103

The obsession with protecting access to lyrics is one of the strangest long-running legal battles to me. I will skip tracks on Spotify sometimes specifically because there are no lyrics available. Easy access to lyrics is practically an advertisement for the music. Why do record companies not want lyrics freely available? In most cases, it means they aren't available at all. How is that a good business decision?

The content industries should have been the ones to invent LLMs, but their head is so stuck in the past and in regressive thinking about how they protect their revenue streams that they're incapable of innovating. Publishing houses should have been the ones to have researchers looking into how to computationally leverage their enormous corpus of data. But instead, they put zero dollars into actual research and develo…

The only people seeing themselves as "content creators" are people giving social media stuff so their users get something they can doom scroll. Other people see themselves as artists, entertainers, musicians, authors, etc.

Re: OpenAI may not use lyrics without license, German court rules

#203

Post by saghm »

input_sh wrote: post #178
saghm wrote: post #143

Earlier quoted context omitted.

One amusing part of lyrics on Spotify to me is how they don't seem to track which songs are instrumentals or not and use that to skip the message about them not knowing the lyrics. An instrumental will pop up and it will say something like "Sorry, we don't have the lyrics to this one yet". The only thing funnier than that is when they do have the lyrics to a song that probably doesn't need them, like Hocus Pocus by F…

Oh they track that, it's in their API as the "instrumentalness" score: https://developer.spotify.com/documentation/web-api/referenc... The fact that they don't do anything with that information is unrelated.

Interesting, especially that it's a probability rather than a boolean! The line can be blurry sometimes (like in the example I mentioned), so it makes sense that it might not be possible to come up with a consistent way of classifying them that everyone would agree with.

Re: OpenAI may not use lyrics without license, German court rules

#204

Post by freejazz »

slaymaker1907 wrote: post #147
freejazz wrote: post #113

Earlier quoted context omitted.

It's a good decision because it must be an incredible minority of people who only listen to music when the lyrics can be displayed. I'd imagine most people aren't even looking at the music playing app while listening to music. Regardless, they are copyrighted and they get license fees from parties that do license them and they make money that way. Likely much more money than they would make from the streams they are…

I think it depends on the music. Most people will have a greatly improved experience when listening to opera if they have access to (translated) lyrics. Even if you know the language of an opera, it can be extremely difficult for a lot of people to understand the lyrics due to all the ornamentation.

What percentage of streaming income does opera, as a genre, represent such that it could even factor into this business decision?

Re: OpenAI may not use lyrics without license, German court rules

#205

Post by terminalshort »

miltonlost wrote: post #112
loudmax wrote: post #68

Simon Willison had an analysis of Claude's system prompt back in May. One of the things that stood out was the effort they put in to avoiding copyright infringement: https://simonwillison.net/2025/May/25/claude-4-system-prompt... Everyone knows that these LLMs were trained on copyrighted material, and as a next-token prediction model, LLMs are strongly inclined to reproduce text they were trained on.

All AI companies know they're breaking the law. They all have prompts effectively saying "Don't show that we broke the law!". That we continue to have tech companies consistently breaking the law and nothing happens is an indictment of our current economy.

This is incorrect. Two judges have now ruled that training on copyrighted data is fair use. https://www.whitecase.com/insight-alert/two-california-distr...

Re: OpenAI may not use lyrics without license, German court rules

#206

Post by riazrizvi »

EdgeExplorer wrote: post #103

The obsession with protecting access to lyrics is one of the strangest long-running legal battles to me. I will skip tracks on Spotify sometimes specifically because there are no lyrics available. Easy access to lyrics is practically an advertisement for the music. Why do record companies not want lyrics freely available? In most cases, it means they aren't available at all. How is that a good business decision?

Thoughts by someone who doesn’t make a living by songs?

I’m guessing you’d want to restrict lyrics to encourage more plays of the song by people who are motivated to understand them. Along with the artist’s appreciation of that experience of extracting what you’re fascinated by. Burdensome processes generate love and connection to things.

Not everything is a functional commodity to be used and discarded at whim.

Re: OpenAI may not use lyrics without license, German court rules

#207

Post by terminalshort »

dahart wrote: post #179
Workaccount2 wrote: post #170

Earlier quoted context omitted.

Training on copyright is not illegal. Even in the lawsuit against anthropic it was found to be fair use. Pirating material is a violation of copyright, which some labs have done, but that has nothing to do with training AI and everything to do with piracy.

There is US precedent for training being deemed not fair use. https://www.dglaw.com/court-rules-ai-training-on-copyrighted... Why wouldn’t training be illegal? It’s illegal for me to acquire and watch movies or listen to songs without paying for them*. If consuming copyrighted material isn’t fair use, then it doesn’t make sense that AI training would be fair use. * I hope it’s obvious but I feel compelled to qualify…

That interpretation is not correct. The owner explicitly denied license to the data and then the company went to a third party to gain access to the data that they were denied license to.

> When building its tool, Ross sought to license Westlaw’s content as training data for its AI search engine. As the two are competitors, Thomson Reuters refused. Instead, Ross hired a third party, LegalEase, to provide training data in the form of “Bulk Memos,” which were created using Westlaw headnotes. Thomson Reuters’s suit followed, alleging that Ross had infringed upon its copyrighted Westlaw headnotes by using them to train the AI tool.

Re: OpenAI may not use lyrics without license, German court rules

#208

Post by terminalshort »

blibble wrote: post #149
miltonlost wrote: post #112

Earlier quoted context omitted.

All AI companies know they're breaking the law. They all have prompts effectively saying "Don't show that we broke the law!". That we continue to have tech companies consistently breaking the law and nothing happens is an indictment of our current economy.

and training on mountains of open source code with no attribution is exactly the same the code models should also be banned, and all output they've generated subject to copyright infringement lawsuits the sloppers (OpenAI, etc) may get away with it in the US, but the developed world has far more stringent copyright laws and the countries that have massive industries based on copyright aren't about to let them evapora…

No thank you. I am perfectly fine with AI training on my open source code and it is perfectly legal because my open source code does not include a license that bans AI training.

Re: OpenAI may not use lyrics without license, German court rules

#209

Post by blibble »

terminalshort wrote: post #208
blibble wrote: post #149

Earlier quoted context omitted.

and training on mountains of open source code with no attribution is exactly the same the code models should also be banned, and all output they've generated subject to copyright infringement lawsuits the sloppers (OpenAI, etc) may get away with it in the US, but the developed world has far more stringent copyright laws and the countries that have massive industries based on copyright aren't about to let them evapora…

No thank you. I am perfectly fine with AI training on my open source code and it is perfectly legal because my open source code does not include a license that bans AI training.

which license is that then?

because other than public domain they all require at least displaying the license, which "AI" ignores

Re: OpenAI may not use lyrics without license, German court rules

#210

Post by qustrolabe »

loudmax wrote: post #68

Simon Willison had an analysis of Claude's system prompt back in May. One of the things that stood out was the effort they put in to avoiding copyright infringement: https://simonwillison.net/2025/May/25/claude-4-system-prompt... Everyone knows that these LLMs were trained on copyrighted material, and as a next-token prediction model, LLMs are strongly inclined to reproduce text they were trained on.

post trained models strongly inclined to pass response similar to what got them high RL score, it's slightly wrong to keep thinking of LLMs as just next token predictions from dataset's probability distribution like it's some Markov Chain
Post reply on HN