Has "the AI did it" ever been a valid defense? I can see why the legislature might want to make it explicit that it is not, but surely it can't be the case that harm caused by an AI is treated as if it were an act of God.
Not yet probably, but "too expensive to comply" is a common refrain from big tech when they break the law. Look at all the excuses Uber has made
I agree. If you hire someone, and they aggrieve someone in the course of their duties, your contractor is your agent and you're responsible for what they did. It makes no sense to treat AI any differently.
While it's not been wrung out in the courts yet, the insurance industry has been denying coverage based on "the AI said so" for a while, coupled tightly with "we can't tell you why because the algorithm is a trade secret". Given the current governance and oversight climate, I don't have much hope that more transparency is in the offing.