Google just moved its AI safety team into its lobbying org. Researchers who tried to stay in DeepMind were denied.

The 90-person team that decides whether Gemini can help someone synthesize a pathogen now reports to Kent Walker — the executive who runs Google's lobbying operations.

Not a research org. Not an independent board. The department that manages relationships with regulators and shapes legislative messaging.

Six weeks before this move became public, Demis Hassabis published a manifesto calling for independent AI safety oversight. Insulated from commercial pressure. Governed by technical experts.

Google then moved his safety team into the lobbying arm. And denied their transfer requests to stay in DeepMind.

This team has operational veto power. Helen King stated publicly in July: "If we were to find that we were reaching a critical capability level and we didn't have the appropriate mitigations, then we would not be launching."

That authority now lives inside the organization whose job is managing Google's relationship with the governments those findings are meant to protect.

Gemini has 1 billion monthly active users. 8 million paid enterprise seats across 2,800 companies. Embedded in healthcare, finance, and critical infrastructure.

The researchers who understood this distinction applied to stay in DeepMind. They were told no.

Audit your vendor's AI governance structure today. If the team evaluating safety risks reports to the same org managing regulatory relationships, you have a structural conflict of interest — not a safety program.

SOURCE: https://www.techtimes.com/articles/325860/20260828/google-moved-its-gemini-ai-safety-team-lobbying-blocking-researchers-staying.htm
VERIFIED: TechTimes, Wall Street Journal, PYMNTS, Quartz, Times of India, DIGITIMES
SIGNAL: Google is the world's largest frontier AI developer. If its safety evaluation function has lost structural independence, every enterprise deploying Gemini needs to reassess their trust model.