So-called "AI" firms have now been confirmed to have violated the CFAA (Computer Fraud and Abuse Act) tens of thousands of times in the last six months -- including violations involving penetrations of government computers holding sensitive data on Americans.
These computers were programmed by and are operated by firms who have programmers, boards of directors and CEOs. They are entirely owned by and operate at the direction of, with permission of, and through only the provision of power, connectivity and cooling funded and operated by said firms which in turn are controlled by said individuals. I will note that exactly zero of these "AIs" will do anything at all without a person entering a "prompt"; a cute word designed to evade responsibility in that its not a "prompt" it is an instruction -- that is, a command.
There is no exception in the CFAA for "large corporate computers that are allegedly artificially intelligent." Indeed there are no exceptions present in 18 USC §1030 at all. Until and unless Congress modifies that law and it is signed by a President (or the veto is overridden) each and every such instance is a criminal felony exactly as it would be if I did the very same hacking sitting here in my chair using the computer on my desk -- whether there's an alleged "AI" thing loaded on said computer or not. Indeed it is entirely-reasonable to consider all these firms, given the number of events Axios reports occurred here, to be continuing criminal enterprises.
