While China endorses open-source AI models, Demis Hassabis heads to Washington to push AI vetting proposal and David Sacks criticized regulators in latest tweet
https://www.bloomberg.com/news/articles/2026-07-16/deepmind-ceo-to-lobby-washington-on-plan-for-group-to-vet-ai-models paywalled but a quick summary and my opinion
Summary of the article:
- Demis Hassabis has been developing this AI vetting proposal for months, consulting with major AI leaders, most notably Dario Amodei, whom he considers a close friend.
- According to Bloomberg, he plans to lobby Washington policymakers next week to gather support for the vetting framework.
- Because Hassabis has a strong reputation and is a Nobel laureate, he is uniquely well-positioned to lead this kind of thing in DC.
My opinion on this: While I understand where Demis is coming from given his long-standing focus on safety, GDM has a track record of rolling back safety mechanisms when it comes to what their AI can and cannot be used for. It’s also questionable how much final say Demis actually has within the broader Google corporate structure.
The timing of this is interesting. It all comes in the same week that:
- The Chinese president openly endorsed open-source AI models (clearly for strategic incentives).
- Moonshot AI dropped a new model that seems just as capable as Fable or ChatGPT on certain tasks.
Meanwhile, David Sacks, Trump’s AI advisor, just criticized these exact types of regulations today. I'm really wondering where all of this leads.