Google Names New DeepMind Chief to Close AI Gap
Koray Kavukcuoglu takes over as senior VP of DeepMind, inheriting pressure to catch up with OpenAI and Anthropic in frontier AI. Meanwhile, Gemini expands…
Koray Kavukcuoglu is now head of DeepMind, Google's AI unit, reporting directly to CEO Sundar Pichai. He inherits one of the tech industry's toughest mandates: catching up to OpenAI and Anthropic in frontier AI development.
Kavukcuoglu, formerly DeepMind's chief technology officer and Google's chief AI architect, replaces founder Demis Hassabis, who moves into the chair role. He will oversee Gemini model development, frontier research, and the Gemini app and developer teams.
The Performance Gap
Google's timing is urgent. The company hasn't released a frontier model since early 2026, while competitors have pulled ahead. Anthropic's Mythos model and OpenAI's GPT-5.6 both broke new ground in recent months, drawing both praise and concern over their advanced capabilities.
Google did release Gemini 3 in November 2025, a model that analysts said "moved the frontier forward" with capabilities "far exceed[ing] what we've typically come to expect from this generation of frontier models." But the momentum has shifted.
"The goal will undoubtedly be to close the gaps with Anthropic and OpenAI in some areas," said Ben Wood, chief analyst at CCS Insight. Kavukcuoglu's promotion "feels like a move that is designed to adjust the focus of Google's efforts away from academic projects and more toward a stronger focus on improving frontier performance and improving the toolset for developers."
Gemini Expands Into Daily Tasks
While DeepMind chases raw capability, Google is embedding Gemini across its consumer products. The new Pixel 11 series showcases a shift toward what Google calls "agentic AI"—systems that act on users' behalf.
U.S. Pixel users can now use Gemini to order groceries, book rides, or call businesses to make reservations or schedule appointments. The AI can handle these tasks autonomously, though users can take over or stop them at any time and review transcripts of AI-operated calls.
New Pixel features also include live translation across videos, podcasts, and voice messages; a transcription tool similar to Wispr Flow; and an accessibility feature that uses Gemini to translate sign language into text. Google has also expanded its gesture-based "Circle to Search" feature to the camera app.
Productivity Tools and User Friction
Google is pushing Gemini into productivity software as well. Google Docs is rolling out a large Gemini bar at the bottom of the screen for some users, a rollout that began in April and continues unevenly across the user base.
Not all users welcome the change. Tech companies are aggressively embedding AI features—adding buttons, pop-ups, and prompts—to drive adoption. For those who want to disable Gemini in Google Docs, the process is straightforward for personal Google accounts: click Gemini in the menu bar, hover over "Bottom bar preferences," and select "Turn off." Fully removing AI features requires either adjusting settings in Gmail or installing a Chrome extension.
For users with Google Workspace accounts managed by employers, the option to disable AI features rests entirely with IT administrators—individuals have no direct control.
Kavukcuoglu faces competing pressures: advancing DeepMind's models while Google simultaneously floods its consumer and productivity layers with Gemini-powered features. The strategy signals confidence that volume and integration can drive adoption even as rivals claim superior technical performance.
