Raw LLM Responses

@L.O.V.E.M.U.F.F.I.N This is exactly the issue. The AI is not the issue because it would still be an issue if it wasn't an AI. If a security process allows a person to look at a Doritos bag and say "That's a gun" and then call the cops to the location without anyone else having to confirm they also saw a gun, or have the person show anyone else what they flagged, that is a problem with the system and not the fact they have AI in it. Programmers will sometimes use AI to assist them in writing code. I definitely do this. The problem is not that, it's when people submit code and say "The AI helped me so if it's wrong it's the AI's fault.". No. YOU submitted the code. Whether you wrote it yourself, you had a human help you, or you had an AI help you, as the person submitting the code YOU are responsible for ensuring the code isn't complete ass. And then after you submit it, multiple other checks, whether done by people or automated systems should also be done. Deflecting blame is the problem. It's the same problem whether you're deflecting blame to another person, or to AI. Not having multiple checks in place is the problem. Blaming AI for causing an arrest because of flagging something that nobody else checked is like blaming a powertool for a building collapsing. If a construction worker is using a nail gun and somehow some of the times he pulls the trigger, no nail comes out, is it the nail gun's fault the structure fails? Or is it the fault of the multiple humans on the team, including the person operating the nail gun, whose job it is to check these things?