According to CoinMeta, Google plans to transfer a team of around 90 people responsible for AI out of DeepMind at the beginning of September and incorporate them into the global affairs department responsible for lobbying and public policy. This is part of the further integration of DeepMind into Google's system. The team responsible for AI mainly studies cutting-edge AI risks, including the chemical, biological, radioactive, and nuclear risks of testing models, as well as the psychological impact of user interactions with AI and chat robots. Some employees are concerned that being moved out of DeepMind will increase the distance from the model development team at Gemini, which may weaken their ability to conduct independent research and identify risks in cutting-edge models. Google stated that the team will still retain access to DeepMind, computing power, and personnel resources, and other teams responsible for model behavior, privacy, and security will also continue to remain in DeepMind. This adjustment comes at a time of ongoing changes in DeepMind management and talent.