Google Reorganizes AI Research Division Amid Concerns Over Safety

Google is reorganizing its AI research and development division, Google DeepMind. Starting in September, the company will move around 90-person AI responsibility team out of DeepMind and into the global affairs organization. This team focuses on emerging and frontier issues in responsible AI, including testing Google's AI models for risks and studying how people interact with them.

The reorganization aims to bring AI safety and responsibility work closer together across the company. However, some employees are concerned that this move will reduce their access to DeepMind researchers and information about new AI models. Helen King, a vice president at Google DeepMind, reassured employees that the team's day-to-day work and main mission won't change.

The team will retain access to DeepMind's computing resources and headcount. The move is part of a broader shift in how DeepMind is organized, with other teams being moved out of DeepMind and into central Google. This reorganization has raised concerns about the potential impact on AI safety research and the team's ability to identify safety questions linked to new models.