Transparency and public trust in METR’s pilot risk report
The author of the report’s AEF-1 disclosure says transparency about limitations and potential biases mattered to them—and that colleagues agreed.
TLDR
The author of the AEF-1 disclosure in METR’s pilot Frontier Risk Report says they valued transparency about limitations and potential biases, and that colleagues agreed. They argue that AI risk assessors seeking public trust should give people the truth “without spin or BS.”
Combined views
22
1 Source, first seen ago