On August 27, as reported by The Wall Street Journal, Google is set to relocate an AI responsibility team, comprising around 90 individuals, from DeepMind to its Global Affairs division—a department dedicated to lobbying and public policy—in early September. This strategic move underscores DeepMind's deeper integration into Google's corporate framework. Concurrently, teams specializing in human resources and policy will also be transitioned out of DeepMind. The AI responsibility team's core focus lies in scrutinizing the risks associated with cutting-edge AI technologies. This encompasses evaluating models for potential chemical, biological, radiological, and nuclear threats, as well as assessing the psychological ramifications of user interactions with AI systems and chatbots.
Some team members have voiced concerns that the increased physical and organizational distance from the Gemini model development team post-relocation might hinder their capacity to conduct independent research and effectively pinpoint risks linked to state-of-the-art models. In response, Google has assured that the team will maintain unrestricted access to DeepMind's resources, including its computational capabilities and personnel expertise. Furthermore, other teams tasked with overseeing model behavior, privacy, and security will continue to operate within DeepMind. This restructuring is taking place against the backdrop of ongoing managerial and talent-related transformations at DeepMind.