Hacker News new | ask | show | jobs
by 0xDEAFBEAD 6 days ago
An AI which is mistaken about the goal you give it, or how to go about achieving it, will behave in a de facto malicious manner, as this incident illustrates.