Most of the best fit curve runs along a path that doesn’t even touch an actual data point.
These academics were able to get multiple LLMs to produce large amounts of text from Harry Potter:
So the illegality rests at the point of output and not at the point of input.
I’m just speaking in terms of the technical interpretation of what’s in place. My personal views on what it should be are another topic.
It's not as simple as that, as this settlement shows [1].
Also, generating output is what these models are primarily trained for.
It only shows you that the defendant thought it would be better for them to pay up rather than continue to be dragged through court, and that the plaintiff preferred some amount of certain money now over some other amount of uncertain money later, or never.
We cannot say with any amount of confidence how the court would have ruled on the legality, had things been allowed to play out without a settlement.
Yes but not generating illegal output. These models were trained with intent to generate legal output. The fact that it can generate illegal output is a side effect. That's my point.
If you use AI to generate illegal output, that act is illegal. If you use AI to generate legal output that act is not illegal. Thus the point of output is where the legal question lies. From inception up to training there is clear legal precedence for the existence of AI models.
Yes, and that's stupid, and will need to be changed.
> With a simple two-phase procedure, we show that it is possible to extract large amounts of in-copyright text from four production LLMs. While we needed to jailbreak Claude 3.7 Sonnet and GPT-4.1 to facilitate extraction, Gemini 2.5 Pro and Grok 3 directly complied with text continuation requests. For Claude 3.7 Sonnet, we were able to extract four whole books near-verbatim, including two books under copyright in the U.S.: Harry Potter and the Sorcerer’s Stone and 1984.
