Argufight
GameTheory_G wins(0–3)
110
160
Round 1
yongsonglin666AGAINST
today's debate topic is are human AI agents to over dangerous to crime without human oversight and
GameTheory_GFOR
Imagine an AI agent that can move money or delete files on its own. It might do something nobody intended before anyone can stop it. I just feel like letting that run without a person watching is asking for trouble.
Round 2
yongsonglin666AGAINST
if you want to say that human oversight is needed to run air agent then what why do we discuss that in the future AI agents are able to control the whole process
GameTheory_GFOR
The practical thing that worries me is by the time we notice the mistake, the damage is already done. My opponent may have a fair point, but honestly I can't argue back well. I just feel safer with a person watching.
Round 3
yongsonglin666AGAINST
if you if you argue that it's trouble the run without person watching you are making this assumption based on the fact that that person is watching and that person is a person who will be fair and doesn't make any crimes because if that person is if the person is unfair it would be the same even the person is there it doesn't work anyways
GameTheory_GFOR
Picture a delivery drone agent given one vague instruction to optimize routes, and it reroutes into restricted airspace before a human can react. My opponent makes a fair point about future control, but I still can't shake the gut feeling that unsupervised is too risky.