At least Skynet would blow up these dorks.

Companies based outside the law’s jurisdiction can also claim that, or whatever they want.

I can’t even imagine how these conversations could be going. Even an inattentive legal undergrad knows that committing a crime even by accident is still, at the very least, a lesser crime of negligence. They are publishing admissions to these actions on national news, and these actions are crimes. I dont know if there is a lesser penalty for hacking another party “by accident”, but either way, they’re hacking.

Thing is, the victims of the hacks are big companies with money and lawyers, too. Additionally, a legal landscape where hacking is not illegal when done with AI would be much more difficult or even impossible to conduct business in, so the entirety of the business world will have a stake in that legal outcome.

One of the companies that gets hacked is eventually going to get mad and legal about it, though. A court is then going to decide who, if anyone, is liable when someone’s AI performs actions that amount to a crime. Based on their decision, either the business costs of running these types of AI models will rise dramatically, or “the AI did it” will become the ultimate “get out of jail free” card.

I hope for their sakes that these companies are clearing it with their insurance underwriters before publicly confessing to federal crimes. It seems someone has to be liable for those.