Google(GOOG.US,GOOGL.US)Gemini AI Intelligent Agent to be Deployed at the U.S. Pentagon for Automating Administrative and Planning Tasks

American tech giant Google (GOOG.US, GOOGL.US) is accelerating the integration of artificial intelligence technology into the U.S. military system. Department of Defense officials stated that Google’s AI agents will be deployed in the work environments of approximately 3 million military and civilian personnel at the Pentagon, used to automate a large volume of routine administrative and operational tasks.

Emil Michael, Deputy Secretary of Defense for Research and Engineering, said that Google’s Gemini AI agents will initially operate on unclassified networks. These AI agents can autonomously perform tasks set by users, such as organizing meeting notes, preparing budgets, and assisting with decision analysis.

In an interview, Michael said, “We are starting with unclassified networks because the vast majority of users work there, and we will gradually expand to classified and top-secret networks.” He also revealed that the Department of Defense is currently discussing the possibility of deploying AI agents in classified cloud environments with Google.

Google stated that Defense Department personnel will be able to create their own AI assistants using natural language, without coding. These digital assistants can automatically perform multi-step repetitive administrative tasks, significantly improving efficiency.

In recent years, the U.S. military has accelerated the application of artificial intelligence. As AI technology becomes increasingly important in military operations, the Pentagon has partnered with several tech companies, including OpenAI, xAI, and Google.

During recent conflicts with Iran, the U.S. military has used AI to assist in identifying military targets and speeding up decision-making processes, greatly enhancing operational efficiency.

However, expanding AI use in the military has also sparked controversy. Previously, American AI company Anthropic had a public conflict with the Pentagon. The company called for restrictions on AI technology use, such as banning its use for large-scale surveillance of U.S. citizens or fully autonomous weapons systems. In response, the Department of Defense listed the company as a supply chain risk last week, and Anthropic subsequently filed a lawsuit against the decision.

While relations between Anthropic and the government have worsened, the Pentagon is rapidly expanding collaborations with other AI companies. Previously, Anthropic was the only AI supplier capable of operating within the Pentagon’s classified cloud system. Now, the Department of Defense has agreements with OpenAI and xAI to run their AI systems in restricted network environments.

Google’s history of military collaborations has also caused internal controversy. In 2018, thousands of Google employees protested the company’s involvement in the Pentagon’s “Project Maven,” which used AI to analyze drone video data. Google subsequently chose not to renew the contract. However, in recent years, the company has gradually eased restrictions on military cooperation.

According to Pentagon plans, Google will initially provide eight preset AI tools to automate tasks such as meeting minutes, budget preparation, and policy review. Some AI tools may also be used for military mission planning and resource assessment.

Since the launch of the GenAI.mil AI platform in December last year, about 1.2 million Department of Defense employees have used Google AI chatbots for non-classified work. Data shows that users have submitted 40 million commands and uploaded over 4 million files.

However, officials also acknowledged that AI training progress still lags significantly behind its usage scale. Since December last year, only 26,000 people have received AI training, and all future training courses are fully booked.

U.S. military personnel stated that AI tools have shown potential in improving military planning efficiency. For example, the U.S. Army’s 18th Airborne Corps designed a military exercise involving 50,000 simulated soldiers, which traditionally took about six months to plan. With the AI platform, similar-scale exercise planning can be completed in just six weeks.

Michael said that although AI may produce errors such as “hallucinations,” risks can be reduced through training, policies, and human review. He pointed out that the Department of Defense will continue to expand AI applications in the future to enhance overall efficiency and combat capability.