Ethan Mollick: governments should directly evaluate AI models instead of trusting third parties

emollick · x · 2026-09-20

Ethan Mollick points out that a report claiming AI gives bad financial advice came from a company selling competing AI financial services products, with questionable methodology that contradicts other recent research. He argues governments need to do direct evaluations of models and publish results rather than trusting third-party evaluators with their own agendas.

Related event: FT claims AI chatbots mostly wrong on finance; Mollick questions methodology(3 posts)→

Original post →

More from Models

Models channel →