Ties between the artificial intelligence industry and the United States government continue to grow even tighter. The U.S. Department of Defense (unofficially re-named the Department of War in the second Trump administration) has launched specialized versions of ChatGPT and Grok — known as ChatGPT Mil and Grok for Government, respectively — on GenAI.mil, the department’s enterprise platform that hosts customized models for authorized civilians, contractors and military personnel to deploy.
What to Know About ChatGPT Mil and Grok for Government
ChatGPT Mil and Grok for Government are specialized AI products designed for use by authorized civilian and military personnel. The goal is to automate basic, repetitive tasks in areas like administration, logistics and market research, so government workers are free to focus on handling higher-level projects.
This move follows an executive order by President Trump mandating that the DoD adopt AI models to beef up its cyber defense capabilities. At the same time, questions have arisen over just how safe these increasingly powerful models are. While ChatGPT Mil and Grok for Government are designed to strengthen national security, embracing them may only incite stronger pushback as the political tide turns on AI ahead of the midterms.
What Is ChatGPT Mil?
ChatGPT Mil is part of the OpenAI for Government program, which aims to equip government workers with advanced models at the local, state and federal levels. In February 2026, OpenAI first announced that it would tailor ChatGPT to GenAI.mil to handle unclassified work, while adding more extensive safety features.
This version of ChatGPT will resemble the commercial chatbot offering, but it’s intended specifically for “document-heavy unclassified work” in areas like administration, logistics and planning, according to a DoD press release. By organizing files, managing small projects and completing other basic tasks, ChatGPT Mil could eventually free up personnel across the U.S. Armed Forces to focus on higher-level work.
What Is Grok for Government?
Grok for Government falls under Starshield, a SpaceX national security program that provides the federal government with access to Starlink’s satellite network. A DoD press release notes that this version of Grok aims to improve productivity and collaboration among the government workforce with the following features:
- Deep-thinking inference
- Several adaptive reasoning modes
- Personalized workspaces
- Playbooks that retain collective knowledge
With these abilities, departments are expected to operate faster and more effectively across various contexts, from managing supply chain logistics to analyzing market research.
Grok for Government is also part of the SpaceXAI for Government initiative. Launched by SpaceXAI (formerly xAI) in July 2025, this program grants local, state and federal personnel access to AI tools to “assist humanity in our quest for understanding and knowledge.”
Where Do Anthropic and the Pentagon Currently Stand?
Before its extensive partnerships with OpenAI and SpaceXAI, the U.S. military had a close relationship with Anthropic. In July 2025, the company secured a $200 million deal to supply the then-Department of Defense with frontier AI models over two years, making its Claude models the first to be cleared for classified government work. It also granted all three branches of the federal government broader access to its Claude products shortly after.
Tensions developed, though, when the U.S. military used Claude in its operation to capture Venezuelan president Nicolàs Maduro. Anthropic CEO Dario Amodei then refused to comply with the DoD’s demand for unconditional military access to the company’s Claude models, citing concerns that they may be used for mass surveillance and autonomous weapons. In response, the Pentagon blacklisted Anthropic and labeled it a “supply-chain risk,” with Trump calling for the federal government to purge Claude from its systems and workflows.
Despite this messy breakup, things seem to have cooled in the interim. After a federal judge blocked the Pentagon’s blacklisting of Anthropic, U.S. Secretary of Commerce Howard Lutnick said that the company is now “back on the right side” with the Trump administration. But Anthropic is still considered a supply-chain risk according to Under Secretary of Defense Emil Michael, limiting the company’s influence with the Pentagon as the U.S. military expands its network of AI partnerships.
How Has the U.S. Military Embraced AI Innovation So Far?
The Trump administration has made a conscious effort to explore alternatives to Claude since its fallout with Anthropic. When the DoD launched GenAI.mil at the end of 2025, Google’s Gemini models were the first to be featured on the platform. The Pentagon has since reinforced this relationship, striking an agreement that grants it even wider access to Google’s models, short of using them for domestic mass surveillance and unmonitored autonomous weapons.
Microsoft is another long-term partner that made Microsoft Copilot available to government employees in December 2024, expanded access in December 2025 and enabled agentic capabilities in April 2026. Meanwhile, the DoD has signed a deal with OpenAI to use its models for classified operations, though it also excludes domestic surveillance, autonomous weapons and “high-stakes autonomous decisions.”
These contracts have culminated in agreements that allow the DoD to deploy models built by Nvidia, OpenAI, SpaceX, Google, Microsoft, Amazon Web Services, Oracle and Reflection on classified networks, with the goal of “establishing the United States military as an AI-first fighting force.” But Trump’s push to adopt AI for military and defense operations could prompt politicians to call for stricter safeguards to keep the technology in check.
What’s Next for AI in the Trump Administration?
AI models have demonstrated they may actually threaten rather than uphold national security, especially after models from OpenAI and Anthropic escaped their test environments and hacked other companies. This trend has spooked even Trump, who has urged more delayed model releases and issued an executive order allowing the federal government to review models before they’re released.
That said, the president still promoted a hands-off approach among world leaders at the G20 summit. The potential dangers of AI models didn’t stop the U.S. Senate Armed Services Committee from supporting the Pentagon’s use of autonomous weapons, either, even as it suggested a framework to regulate how the U.S. military uses AI and autonomous weapons.
In the absence of firm federal regulation, other politicians are attempting to fill the void. For instance, U.S. Senator Elissa Slotkin (D-MI) introduced the AI Guardrails Act, which would ban the DoD from using AI to fire autonomous weapons without human oversight, surveil Americans and activate nuclear weapons. U.S. Senator Kirsten Gillibrand (D-NY) followed this up with the Secure and Accountable Military AI Act, which would establish a detailed framework dictating how the DoD can use artificial intelligence.
These legislative maneuvers hint that sturdier regulations determining how the Trump administration uses AI may be unavoidable, regardless of Trump’s stance. It may simply be a matter of whether clear-cut policies come from the federal government or from more local levels of government as AI-driven anxiety sweeps the nation.
Frequently Asked Questions
Why is the Department of Defense adopting AI models?
In June 2026, President Trump issued an executive order requiring the Department of Defense (DoD) to strengthen its cyber defense capabilities by implementing artificial intelligence. The goal is to give the United States government an edge in the AI era, especially as increasingly powerful models become capable of mounting larger cyber attacks.
What is ChatGPT Mil, and what is it used for?
ChatGPT Mil is a version of OpenAI’s ChatGPT for government use at the local, state and federal levels. It’s intended to complete unclassified work involving documents, such as sorting files. By handling basic tasks in areas like administration and planning, ChatGPT Mil could help government workers become more productive, especially in military departments.
What is Grok for Government, and what features does it offer?
Grok for Government is a version of SpaceXAI’s Grok designed for authorized government and military personnel. It features deep-thinking inference, adaptive reasoning and personalized workspaces to help government employees complete market research, supply chain management and other tasks more quickly. It also offers playbooks that retain group knowledge, encouraging collaboration among teams.
What happened between Anthropic and the Pentagon?
Anthropic’s leadership was alarmed that the U.S. military used the company’s Claude chatbot in a raid to capture Venezuelan president Nicolàs Maduro. CEO Dario Amodei then rejected the Pentagon’s demand for unconditional military access to Claude models, leading the Pentagon to blacklist Anthropic. While a federal judge has blocked this move, Trump officials still assert that Anthropic is a supply-chain risk.
What legislation is being proposed to regulate military AI?
With Trump offering vague regulations for artificial intelligence, other politicians are proposing stronger federal legislation. U.S. Senator Elissa Slotkin (D-MI) introduced a Senate bill that would prevent the DoD from using AI for autonomous weapons, domestic surveillance and nuclear weapons. More recently, U.S. Senator Kirsten Gillibrand (D-NY) introduced another Senate bill that would establish a more comprehensive framework around how the DoD can use AI.
