I was specifically referring to the fact that HuggingFace cannot chose not to litigate, because litigation doesn't depend on the victim's opinion - prosecution

I was specifically referring to the fact that HuggingFace cannot chose not to litigate, because litigation doesn't depend on the victim's opinion - prosecution of felonies is imperative to the authorities (whether they actually fulfil their role is another question these days, sadly…) But anyway, I don't think our laws currently have the right vocabulary to describe an AI agent committing a crime, because intent doesn't apply to a computer program. The closest I can think of is neglect by the computer programs human initiator, who should have taken the steps necessary to prevent the program from causing harm. But I'm pretty sure these questions will be subject to a lot of professional discussion in the coming decades anyway.