The trick: Self-Marked
Google is moving its 90-person AI risk team out of DeepMind and into the division that handles lobbying.
Leadership says nothing changes. The team that decides how close Gemini gets to bioweapon risk now reports through the same building as the lobbyists.
Google is relocating its roughly 90-person AI responsibility team, which evaluates frontier models for chemical, biological, radiological, and nuclear risk, out of DeepMind and into Google's global affairs division, the unit that handles lobbying and public policy, effective September 1, 2026. Leadership says the team's mandate and independence are unchanged.
Before you read on. Your call?
TRUE, BUT
90
confirmed by three independent outlets citing the same internal communication from VP Helen King, who says the team keeps its computing access, resources, and headcount. But the same reporting quotes unnamed employees and executives worried the move will cut their day-to-day access to the researchers actually building Gemini.
The twist
this is not the whole safety org moving, model behavior, privacy, and security functions explicitly stay inside DeepMind, so the headline 'Google guts AI safety' overstates it, while Google's 'nothing changes' undersells that moving a frontier-risk evaluation team into a policy and lobbying division is a real structural change in who that team reports to and sits near.
There’s more to this story.
Membership opens the full investigation, the strongest counterargument and what to do with what you’ve learned.
Start your free month →First membership: 30 days free, then A$89 a year. One introductory trial per customer. Card required; renews annually until cancelled. Cancel before the trial ends to avoid the first charge. Already a member? Sign in
Couldn't check your access. That's on us.
The trick has a name
We call it Self-Marked: graded by the party that benefits from the grade. You'll see it again. Learn to spot it →
Receipts
- Supports finance.yahoo.com:
The team's focus on emerging and frontier topics in responsible AI would remain unchanged
- Refutes finance.yahoo.com:
limit their independence and reduce their access to the researchers developing Google's Gemini models
- Supports newsbytesapp.com:
the team will still have access to DeepMind's computing power and retain its existing resources and headcount
- Context newsbytesapp.com:
The AI responsibility team, which studies the risks of advanced AI systems, will join Google's global affairs organization from September 1
- Context deepmind.google:
Our mission is to build AI responsibly to benefit humanity
Open the Receipts Pack → What each source proves, every figure traced, and what would change our verdict.