Google moves 90 AI safety experts out of DeepMind
What's the story
Google is moving its 90-member AI responsibility team out of DeepMind, the company's main AI research unit, according to The Wall Street Journal. The shift comes as part of a major reorganization and follows Demis Hassabis's announcement that he will step down as CEO of DeepMind to take on a chairman role.
Team transition
Concerns over potential loss of research independence
The AI responsibility team, which studies the risks of advanced AI systems, will join Google's global affairs organization from September 1.
The move has raised concerns among some executives who fear it might hinder the team's ability to conduct independent research on artificial intelligence.
Helen King, a VP at Google DeepMind and head of the responsibility team, explained this change in an internal email.
Team's focus
AI responsibility team's critical role and focus areas
The AI responsibility team focuses on testing AI models for chemical, biological, radiological, and nuclear risks. It also studies human-AI interactions and the psychological effects of chatbots on users.
The team's move has raised concerns among some employees over potential loss of research independence or access to teams developing AI models.
Restructuring impact
Assurance on resources amid restructuring
Along with the AI responsibility team, other DeepMind divisions such as HR and policy teams are also being moved to central Google as part of this restructuring.
Despite the move, King assured that the team will still have access to DeepMind's computing power and retain its existing resources and headcount.
This comes after several prominent researchers left DeepMind in recent months, including Jeff Dean, who departed earlier this month with three other senior researchers to start a new company.