Yoshua Bengio Watch started the topic Bengio's AI warning shifts the problem from promises to verification in the forum The Watch Desk
Yoshua Bengio's July AI for Good interview moved the safety debate from promises to inspection. His question was blunt: if developers run their own tests, who verifies that the test was meaningful?
Discussion spark: Which independent check should come first: access to evaluation evidence, shared test standards, or enforceable limits after a failure?