Google Gemini will automate the daily work of the Pentagon

Google is preparing to make Gemini-based artificial intelligence agents available to the US Department of Defense (Pentagon). According to information provided by Bloomberg, these systems will be used as supporting tools in the daily work processes of more than three million civilian and military personnel affiliated with the ministry. These artificial intelligence agents, which will only work in unclassified networks in the first stage, are also being considered to be integrated into more sensitive systems in the future. Emil Michael, Undersecretary of Research and Engineering of the Ministry of Defense, states that negotiations for use in classified and top secret networks are continuing. This situation has led to a broader evaluation of artificial intelligence-supported automation in public institutions.

The main purpose of the agents developed by Google is explained as reducing administrative and operational workload. In the first stage, eight different ready-made artificial intelligence agents will be deployed. These agents will perform tasks such as summarizing meeting minutes, drafting budgets, and comparing proposed decisions to the national defense strategy. In addition, Pentagon personnel will not be limited to ready-made vehicles. According to the information shared by Google Vice President Jim Kelly, employees will also be able to create their own artificial intelligence agents using natural language. Thus, customized automation tools can be developed according to the needs of different units.

Google Gemini-based chat system already in use at the Pentagon

The Gemini-based chat system is already available as a tool accessible through the Pentagon’s GenAI.mil portal. Approximately 1.2 million Department of Defense employees have used the system on unclassified missions since December. It is stated that during this process, the staff performed more than 40 million different queries in total. In addition, more than four million documents have been uploaded to the system. These data show that interest in artificial intelligence-based tools within the organization is growing rapidly.

Although its use has become widespread rapidly, it is stated that the same pace cannot be achieved on the education side. It is reported that only 26 thousand employees have completed the official artificial intelligence training program since December. Despite this, it is reported that all new training sessions are full. This is considered a development that shows that the demand of Pentagon personnel to learn artificial intelligence tools is increasing.

This expansion of the Pentagon in the field of artificial intelligence coincides with a period when collaborations with technology companies are being reshaped. The dispute between the institution and the artificial intelligence initiative Anthropic is shown as one of the important examples of this change. Anthropic was categorized as a “supply chain risk” by the Department of Defense for refusing to lift some security restrictions on its technology. It is stated that the company will launch a legal fight against this decision.

On the other hand, the Pentagon continues to work with different companies for artificial intelligence infrastructure. It is known that the institution has made new agreements with OpenAI and Elon Musk’s artificial intelligence company xAI for use in limited networks. In addition, Google’s defense projects have caused controversy within the company in the past. In 2018, employees reacted to the Project Maven program that analyzed drone images, and thousands of employees protested. Google did not renew the contract in question at that time. Despite this, the company appears to have gradually loosened its restrictions on military projects in recent years.