Post #4415878

2026-08-06 11:14 UTC

@flyingpenguin@infosec.exchange @Viss@mastodon.social @jfslowik@infosec.exchange It would be an incredibly difficult case to prosecute. You'd have to convince a jury that the decendants' actions meet all the elements of the crime, including poor configurations and vague prompts, none of which explicitly asked or even suggested the model hack into an outside entity's network. IMO we really need new laws that establish crimes based on recklessness, negligence, or failure to supervise autonomous agents.

Replies (1)

@flyingpenguin@infosec.exchange 2026-08-06 11:34

@DaveMWilburn@infosec.exchange @Viss@mastodon.social @jfslowik@infosec.exchange semi-autonomous agent given goals, tools, and latitude. the law's answer for forty years has been more autonomy granted, more supervision owed. Tesla drove this in reverse, shipping autonomy and blaming the driver no matter what, and that worked right up until a jury said no for 243 million bucks. loophole cowboys shouldn't fool anyone. look at Section 15(b)(4)(E) of the Exchange Act, in force since 1964 with sharpened teeth after the 1988 Insider Trading and Securities Fraud Enforcement Act: a broker-dealer or supervisor is liable for violations committed by any person subject to their supervision, if they failed reasonably to supervise with a view to preventing the violation. note what's absent: intent that the trader commit the act. nobody prosecutes Morgan Stanley by proving management asked for rogue trades.

Open ##4415877