Comment by ibeckermayer
10 hours ago
Trusting the output of an abstraction explicitly designed, engineered, and tested by other humans with an actual understanding of the layer beneath you is not comparable to trusting a non-deterministic, statistical next-word-predictor with no mind or ability to be morally responsible for what it produces.
Assuming most humans accept moral responsibility seems rather dangerous in itself. Especially as they climb the corporate ladder.
Luckily, accountability does exist in most professional settings so someone accepting moral responsibility is not required.
It doesn't always land on the right person but it's usually a motivating factor for people to make an effort.