Robert66 wrote: ↑June 15th, 2022, 4:22 pm
Pattern-chaser wrote: ↑June 15th, 2022, 1:47 pm
Robert66 wrote: ↑June 14th, 2022, 5:39 pm
Pattern-chaser wrote: ↑June 14th, 2022, 2:33 pm
Likewise. There are too many innocent people convicted for us to be at ease with such an irreversible punishment. First let's guarantee that no innocent people are convicted; then, perhaps, we can safely re-consider capital punishment...?
Yes, let's! How though?
Robert66 wrote: ↑June 14th, 2022, 6:24 pm
Should we feed all the evidence and arguments into a supercomputer for an objective, flawless techno-judgment, or continue to rely on error-prone humans?
Error-prone humans for me, every time. Do we really think that AIs could offer error-free judgement(s)? Their objectivity and flawlessness depends on the way it is programmed, maybe by humans, maybe reprogrammed by itself. I'll opt for the 'devil we know'.
Sounds to me as though you trust an error-prone system. In that case I agree with you.
No, I'm choosing between 2 undesirable choices. Humans make mistakes, but we have much past experience of the sort of mistakes that are made. In the case of AIs, there is no telling what sort of errors they might make. As to the latter, I assume it is humans identifying the error(s), not the AIs, which might think their conclusions correct, according to the criteria they have adopted.