Google has announced the deployment of Gemini AI agents to over 3 million civilian and military personnel within the U.S. Department of Defense. This marks a significant advancement in Google's collaboration with the Pentagon, signifying the formal integration of mainstream AI technology into the administrative and operational levels of the U.S. core defense system.

Emir Michael, the Under Secretary of Defense for Research and Engineering, stated that these AI tools are currently authorized only for use on non-classified networks, and whether they will be expanded to classified or top-secret systems remains under further discussion.

Functions and application data of the first AI agents:

  • Administrative Automation: The initial eight customized agents focus on handling tedious administrative tasks, including automatically organizing meeting minutes, assisting in departmental budget preparation, and reviewing action plans for compliance with national defense strategies.

  • Self-creation of Permissions: In addition to official tools, DoD staff can also create customized AI agents tailored to specific needs by using natural language instructions.

  • High-frequency Interaction Demands: Since the trial operation began last December, approximately 1.2 million employees have submitted around 40 million prompts through a dedicated portal, processing more than 4 million documents, indicating strong demand for AI tools within the military.

  • Training Lag Challenges: Despite the sharp increase in users, training progress remains slow. Currently, only about 26,000 people have completed the AI training courses for the system, far below the actual number of users.