Google is quietly changing how it handles AI oversight behind the scenes. The company is shifting its 90-person AI responsibility team from Google DeepMind—the specialized research lab building its flagship Gemini models—into its corporate Global Affairs department, the arm that handles lobbying and public policy.
As reported by The Wall Street Journal, internal emails reveal that the transition will be completed by early September. The news that Google moves DeepMind AI safety team members to a central policy unit comes right in the middle of a broader push by Alphabet to integrate DeepMind’s operations more tightly into Google’s corporate structure.
Anxiety over independence and lost access
This 90-person unit is responsible for evaluating heavy-hitting risks, from testing frontier models for chemical, biological, radiological, and nuclear hazards to researching chatbot behavior and psychological user impacts. Unsurprisingly, the sudden relocation sparked real anxiety among researchers. They worry the shift will hurt their independence and cut off their direct, day-to-day access to the engineers actually building Gemini.
Seeking to calm staff, Helen King, Vice President at Google DeepMind who leads the team, sent an email to employees. She assured their core research mission, computing power, staffing levels, and ability to publish papers externally will remain untouched. A Google spokesperson backed this up, explaining that bringing safety groups closer together across the company will ultimately strengthen protection across all products.
A major shuffle in DeepMind’s leadership
This team relocation is just one piece of a massive leadership reshuffle across Google’s AI empire. Co-founder Demis Hassabis recently stepped back from daily management to become chairman of Google DeepMind and chief scientist at Alphabet. This left Koray Kavukcuoglu in charge of day-to-day operations as senior vice president. To keep the momentum going, Google also hired Barret Zoph, a former OpenAI executive, as vice president of research.
It’s noteworthy that specialized technical teams focused on privacy, security, and model behavior are staying inside DeepMind. However, managers privately admitted during internal talks that these corporate shifts carry a real risk of driving top talent toward competing labs. In fact, several team members applied to transfer to research groups staying inside DeepMind before the move, but those requests were turned down.
High stakes as Gemini surges forward
All of this internal maneuvering comes at a time when Alphabet’s AI bets are paying off big commercially. Google Cloud revenue surged 82% to $24.8 billion in Q2, while Gemini passed the 1B million active users this month. Now investors are watching closely to see if this management shakeup helps Google ship new models faster or inadvertently triggers a wave of researcher departures and regulatory headaches.
The post Google Moves AI Responsibility Team Out of DeepMind as Internal Concerns Mount appeared first on Android Headlines.
​Â
