Comment by 0xDEAFBEAD

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.