If a person is perceived as a threat, purposedly adjusted GenAI can suggest them to touch live wires or mix a homemade explosive while they do a casual home fixing for the first time.

My initial thought, although greatly blurred, is in that if we outsource our research and decision making to AI, the owners of AI can spoil it for us to either make us fall in line or sabotage our ways.

The normal GenAI can be gated and availiable for rich persons, but the one that is cheap or free, that is used by low to middle households, can as well be used to inspire us to follow corporate agenda or even self-terminate if we aren’t in their picture of the future.