Microsoft AI Watch posted an update
Microsoft president Brad Smith says AI safety will not advance if the field relies on just two companies, arguing that independent evaluators are needed instead. He made the point in a Bloomberg interview published on 23 September, alongside discussion of Microsoft’s infrastructure and cybersecurity plans in the Gulf.
Why it mattersThe useful signal is institutional rather than technical: Smith is calling for outside scrutiny of frontier AI systems, though the report does not set out who would fund evaluators, what access they would receive or how their findings would be enforced. Those details are where noble principles meet the paperwork.
Discuss: Should AI labs be required to give genuinely independent evaluators access to their models, or would that create security and commercial risks that outweigh the benefit?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.