When AI models disagree on a threat assessment, the disagreement gets logged....

https://simonsnewchat.rivetgarden.com/posts/how-to-create-a-synthetic-workplace-threat-case-for-tabletop-testing

When AI models disagree on a threat assessment, the disagreement gets logged. Each review records its assumptions, names the human who owns the final call, and schedules a follow-up check. The review interval is set per case, not fixed at 30 days