Google Breaks Up AlphaFold Team Behind Nobel Prize AI

Google DeepMind has disbanded the team behind its Nobel Prize-winning AlphaFold AI system, the Financial Times reported Wednesday.

The move reshuffles the research group that produced one of the most celebrated achievements in artificial intelligence. Most of the original authors of the AlphaFold papers have been reassigned over the past year, according to the FT's analysis of job moves and people familiar with the matter.

The company confirmed that the employees moved to projects tied to Google's Gemini large language model, as well as enzyme design and nuclear fusion. Others went to Isomorphic Labs, Google's drug-discovery spinout. Almost a quarter of the full-time Google DeepMind authors of the original AlphaFold papers have left the company entirely.

"Our strategy over the last nine years has been to focus on grand challenges," said Pushmeet Kohli, vice-president of research at Google DeepMind and head of the AI for Science team. "The strategy has evolved."

DeepMind is now concentrating on building Gemini-powered systems that can help scientists and eventually automate parts of the scientific process, Kohli told the FT. That puts it in more direct competition with OpenAI and Anthropic, which are also racing to build general-purpose AI agents.

The reshuffle comes weeks after John Jumper, one of the Nobel-winning scientists behind AlphaFold, left Google for Anthropic. Jumper shared the 2024 Nobel Prize in Chemistry with DeepMind CEO Demis Hassabis for the protein-folding breakthrough.

AlphaFold, which predicts protein structures from amino acid sequences, was released in 2021 and widely hailed as a landmark in biology. The FT's report frames the reorganization as Google pivoting from pure scientific research toward more commercially focused AI products.

In a separate development, Google released research showing that workers in physical and manual roles, such as auto technicians and industrial mechanics, are using conversational AI for real-time diagnostics and troubleshooting. The company's research found that workplace AI now touches 68% of jobs, representing 90% of employment in the U.S. Within any single job, employees use it for about 21% of their tasks on average.

"Blue-collar workers tend to be using a lot of what we call multimodal AI, which is AI with images and video," Scott Strand, head of strategic operations and special projects for technology and society at Google, told Axios last week.