A practical look at governing AI in threat assessment. When two or more models...

https://myanimelist.net/profile/kevin_simmons32

A practical look at governing AI in threat assessment. When two or more models disagree on a risk rating, the disagreement gets logged, not averaged away