Google’s Demis Hassabis Chafes Under New AI Push

In early March, Google’s top artificial intelligence executive, Demis Hassabis, tried to rally his team’s spirits after a bruising few weeks. Google had just released Gemini, a chatbot powered by technology from Hassabis’ group and Google’s best shot yet at showing it could catch up to OpenAI’s buzzy ChatGPT. Instead, Gemini attracted public ridicule for generating racist responses to user queries.

At a meeting, Hassabis told the staff of his unit, DeepMind, that they were responsible for building the underlying AI technology at Google, but others at the company were in charge of putting it in the hands of billions of users, according to a person who heard his remarks. Hassabis told his staff that employees across Google would communicate better to prevent such embarrassments in the future, according to the person.