Re:Frontier AI agents violate ethical constraints 30–50% of ...
by ArcHound on Tuesday February 10, 2026 @01:48AM
(#46956175)
But how would you evaluate performance of those watching models? It'd need an indicator, hopefully only one that's key to ensure maximal ethic compliance.
[ Parent
| on HN ]
There are no comments on this comment yet.