Pentagon expands generative AI to 1.5 million users

What happened?

The Pentagon has announced that its generative artificial intelligence platform, GenAI.mil, has grown from 80,000 to 1.5 million daily users in just six months, according to statements by Deputy Secretary of Defense for Research and Engineering Emil Michael. The platform, based on Google Gemini Enterprise, is available to nearly half of the Department of Defense's (DoD) 3.5 million employees, including civilian personnel, military, and contractors. This growth represents an 18.75-fold increase in daily active users, far exceeding initial expectations. According to The Next Web, the original pilot launched in early 2024 and expanded rapidly after receiving positive user feedback.

Why is it important?

This exponential growth demonstrates that generative AI is being adopted at an unprecedented scale in the government and military sectors. Unlike limited pilot projects, GenAI.mil has become a daily operational tool. The speed of adoption even surpasses that of many tech companies: for example, Microsoft Copilot took over a year to reach 1 million enterprise users. The historical context is key: in 2020, the DoD launched the Joint AI Center (JAIC) with modest projects; now, GenAI.mil represents a qualitative leap. Additionally, the use of AI in defense raises questions about security, ethics, and human control, especially in critical applications such as mission planning or intelligence analysis. The integration with Google Gemini also raises concerns about reliance on an external vendor and potential transfer of sensitive data.

Consequences and context

The expansion of GenAI.mil could accelerate AI integration in other government agencies, such as the State Department or NASA. In fact, the DoD is already sharing lessons learned with other agencies. However, it also raises concerns about potential biases (language models can reflect training data biases), security vulnerabilities (prompt injection attacks, leakage of classified information), and dependence on external vendors like Google. The Pentagon has implemented safeguards such as mandatory human oversight for high-risk decisions, limiting certain uses (for example, GenAI.mil is not allowed for autonomous weapons systems), and periodic audits. Nevertheless, the rapid deployment could outpace existing regulatory frameworks. Compared to other defense AI projects, such as Google's Project Maven (2018) which sparked internal protests, GenAI.mil has been received more positively, but doubts persist. Emil Michael told The Next Web: 'The adoption of generative AI in the Pentagon is the fastest in the history of the U.S. government.'

What readers should know

• GenAI.mil is a generative AI platform based on Google Gemini Enterprise, adapted for military use with additional security layers.
• It currently has 1.5 million daily active users, up from 80,000 six months ago, representing a 1,775% growth.
• It is used for tasks such as report writing, data analysis, intelligence summaries, logistical decision support, and code generation.
• The Pentagon ensures that usage is supervised and restricted to prevent unethical applications; for example, queries about nuclear weapons or offensive cyberattacks are blocked.
• This case could serve as a model for other government agencies worldwide, such as the UK Ministry of Defence or NATO.
• The program's cost has not been disclosed, but the agreement with Google is estimated to be tens of millions of dollars per year.

'The adoption of generative AI in the Pentagon is the fastest in the history of the U.S. government.' — Emil Michael, Deputy Secretary of Defense for Research and Engineering

The original article from The Next Web highlights that the platform initially launched as a pilot in January 2024 and has grown organically through word of mouth among employees. No major security incidents have been reported, but cybersecurity experts warn about risks of data leakage or misuse, especially if employees inadvertently input classified information. The DoD has implemented a system to monitor prompts and responses to detect anomalies. Compared to other government AI implementations, such as the use of ChatGPT in the Biden administration (which was restricted due to security concerns), GenAI.mil represents a more open but controlled approach. The platform also competes with internal solutions like the Air Force's 'Project Salus,' but its massive adoption suggests it has become the de facto standard tool.