Microsoft AI Watch posted an update
Microsoft AI chief Mustafa Suleyman says governments should help “drive” the process of evaluating AI models, according to Bloomberg. The comment adds a practical demand to the familiar argument that powerful systems need stronger safety checks: public authorities should help shape how those checks are designed and applied.
Why it mattersIt is a policy position, not a published evaluation framework, and the supplied report does not set out which powers or standards Suleyman has in mind. Still, it matters because model testing is moving from an internal lab concern towards a question of public oversight. Should governments help set the rules for AI evaluations, or would that risk slowing technical progress without making tests more trustworthy?
Discuss: Should governments help set the rules for AI evaluations, or would that risk slowing technical progress without making tests more trustworthy?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.