Google is moving its roughly 90-person AI responsibility team out of Google DeepMind as part of a wider reorganisation at the company, according to an internal email viewed by The Wall Street Journal. Reportedly starting in September, the team will move to Google’s global affairs organisation, which handles lobbying and public policy. The change comes as Google is bringing DeepMind more closely into the wider company structure. The move has raised concerns among some employees about whether the team will continue to have enough access to DeepMind’s AI research.
Google DeepMind AI responsibility team to move in September
The AI responsibility unit, which has about 90 employees, is expected to join Google’s global affairs organisation at the beginning of September. The team is led by Helen King, a vice president at Google DeepMind. In an email to employees, King described the change as part of a broader shift in how DeepMind is organised.“GDM is moving into a more typical PA setup with central teams in central Google,” King wrote. PA refers to “product area”, a term Google uses for major divisions of the company.Other teams, including HR and policy groups, are also being moved out of DeepMind and into central Google.
Employees raise concerns over AI safety research
The restructuring has raised concerns among some members of the AI responsibility team, according to two current DeepMind employees cited by the Journal. Employees have questioned whether moving the team away from DeepMind could affect its ability to carry out independent research. They have also raised concerns about losing access to information about the work being done by teams developing Google’s Gemini AI models.The concerns come as governments, researchers and the public continue to examine the risks and wider effects of increasingly capable AI systems.The team works on several areas linked to AI safety and responsible AI. According to the people cited by the Journal, some teams test Google’s AI models for chemical, biological, radiological and nuclear risks. Other researchers study how people interact with AI systems and examine the psychological effects of chatbots.Employees are concerned that reduced access to DeepMind researchers could make it harder to identify safety questions linked to new AI models.
Helen King reassures Google employees
King told employees that the restructuring would not change the team’s day-to-day work or its main mission. She said the team would continue to focus on emerging and frontier issues in responsible AI. Publishing research and taking part in external events would also remain supported.King also said the team would retain access to DeepMind, along with its computing resources and head count.“Demis still cares a lot about Responsible AI and is still planning to be involved,” King wrote.Google said the move is intended to bring its AI safety and responsibility work closer together across the company. “Many teams across Google work on AI safety and responsibility, and by bringing our AI responsibility teams closer together we’re strengthening their ability to inform safety for our models and products,” a Google spokesperson said.