Hacker News Reader

CC-BY-SA to be specific. How does attribution of derivative works work in LLMs (etc)? When is the content they produce required to be CC-BY-SA as well (by virtue of the -SA part of the license)?

View on HN

LLM output can't be copywritten as it's not the work of a human. The company would presumably have to include attribution to every piece of data the system is trained on for every query? Seems absurd, but then how would attribution work? Not really considered when the license was written.

Licenses go beyond copyright though, they tell you what you can and can't do with something someone else made. They're contracts.

I'm sure we'll be seeing more lawsuits (and likely new regulation and licenses) around this.

Are you saying LLMs are not derivatives of their source material? Why not?

I agree on lawsuits being likely and needed to establish precedent. CC-BY-SA specifies "other rights such as publicity, privacy, or moral rights may limit how you use the material".

At some point, maybe poorly-trained chatbots that consistently produce what's seen as avoidably/negligantly poor results may become regulated. Like how if a company poorly trains its employees, it is on the hook for their employees' behavior.

Reply on news.ycombinator.com