Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Keep in mind the 4 factors of Fair Use (US-specific):

  1. the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes;
  2. the nature of the copyrighted work;
  3. the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and
  4. the effect of the use upon the potential market for or value of the copyrighted work.
For 1, maybe OpenAI could've been safe if they'd actually stayed "open", but nowadays every AI company clearly fails, as do many (but not all) of the LLM users. Contrast this with most traditional fanfiction and personal projects where there were scary letters and occasional bullying, but few actual law-based problems.

"Transformative" is also part of 1 and is often cited as letting LLMs get away with everything, but everybody argues that and doesn't always win. Also, it's quite linked with 4.

2 mostly isn't a problem but gets into nasty details.

For 3 (emphasis on "used"), this link once again proves that the point does fail.

For 4 we are indisputably seeing mass disruption in several fields, so that point fails.

To be clear - there are ways to use LLMs that balance much closer to the side of fair use, but that isn't how LLMs are advertised.






I feel like once you start talking about a specific copyrighted work this falls apart.

We have not seen mass disruption in books or news for example, which seem to be two big areas that are most aggressively pursuing copyright claims.

I think the best case on this front is probably from Getty where image generation models ARE directly competing with them.




Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: