Google has announced significant changes to its artificial intelligence leadership following the departure of several senior researchers, signaling a shift in its approach amid increasing competition in the AI sector. On Wednesday, Demis Hassabis, the CEO of Google DeepMind and recent Nobel laureate, stepped down from his operational role to assume the positions of chair of DeepMind and chief scientist at Alphabet, Google's parent company. His transition aims to allow him to focus on long-term strategic initiatives, including his work on Isomorphic Labs, an AI-driven drug discovery project.
The announcement came hours after Jeff Dean, a pioneering AI researcher who played a key role in developing Google's foundational AI technologies, revealed he was leaving to start a new venture called Discovery Loop with three other senior researchers, including Quoc V. Le, Oriol Vinyals, and Sanjay Ghemawat. Dean’s exit is viewed by industry observers as a pivotal moment for Google and raises concerns about the company’s ability to retain top talent amid intensifying competition from rivals like OpenAI and Anthropic.
The reshuffle reflects ongoing challenges for Google as it contends with missed opportunities during the AI surge sparked by OpenAI’s ChatGPT launch in late 2022. Though Google responded by merging its AI research units—Google Brain in Silicon Valley and DeepMind in London—under Hassabis’s leadership in 2023 and revamping its flagship search engine to incorporate AI capabilities, the company has struggled to keep pace with its competitors. A highly anticipated release of its Gemini 3.5 Pro AI model, originally expected in June, was delayed, further highlighting development hurdles.
Koray Kavukcuoglu, a DeepMind co-founder and the lab’s chief technology officer, will assume day-to-day operational control of the merged AI group. Kavukcuoglu, who relocated to Google’s Mountain View headquarters last year, will report to both Hassabis and CEO Sundar Pichai. His expertise includes early breakthroughs in reinforcement learning and managing DeepMind’s complex AI infrastructure.
Google has continued to invest heavily in AI, forecasting 2024 expenditures on data centers and AI-related initiatives to reach between $195 billion and $205 billion, more than doubling last year’s $85 billion spend. Despite this, smaller Google AI models released in July, including one focused on cybersecurity, did not outperform competitors on key benchmarks.
Within Google, Dean and Hassabis have been seen as moral compasses, with Dean notably raising ethical concerns about military contracts. Some DeepMind employees have expressed unease that these leadership changes, combined with integration into Google's broader AI division, may reduce the lab’s independence and diminish its focus on ethical boundaries.
Google’s CEO Sundar Pichai praised Hassabis, emphasizing the importance of his new role in shaping AI’s future both for Alphabet and broader society. Hassabis has recently emerged as a leading figure in AI governance discussions, engaging with policymakers and international leaders, including at the recent G7 summit.
As Google navigates these transitions, industry experts and insiders watch closely to see how the company will sustain its innovation and leadership in a rapidly evolving AI landscape.