On December 9, 2025, the Secretary of War issued a memo that should’ve landed on your screen whether you’re an E-4 working logistics, a butter bar in a TOC, or a GS-13 in acquisitions: “The War Department Unleashes AI on New GenAI.mil Platform.”

In short, the Department of War (DoW) approved and launched GenAI: a generative artificial intelligence platform, running on Gemini and certified up to Impact Level 5 (IL5). That means it can process Controlled Unclassified Information (CUI) on the NIPR side of the house. It’s built to assist service members with different tasks.

AI Is a Tool but Under the UCMJ, You’re Still the Operator

But one may ask, what does this have to do with the Uniformed Code of Military Justice (UCMJ)? Well, quite a lot because the UCMJ has zero chill when it comes to negligence, reckless, or unauthorized use of government systems. Military law may not mention “AI” by name, but misuse of AI tools is clearly punishable under several articles of the UCMJ. In other words, there’s no article that says “Thou shalt not use AI.” But trust me, there are plenty of articles that may apply to AI misuse.

GenAI.mil is cleared to handle CUI only, not classified data.

Whether you’re in an ops shop or the motor pool, you need to understand two terms:

1. Classified: Information that has been designated as requiring protection for national security—usually Top Secret (TS), Secret (S), or Confidential (C); and
2. Controlled Unclassified Information (CUI): Sensitive government data that doesn’t rise to the level of classified, but still demands protection under law or policy (think PII, health data, logistics movement, and some operational plans).

If you load anything from SIPR, spill from a top-secret memo, or even paraphrase classified tactics into this system, you may have just crossed into unauthorized disclosure territory. And that, friend, could land you with an investigation and potential Non-Judicial Punishment (NJP) under Article 92. Article 134 could also be in the mix, if the chain of command decides you’re a liability to national security.

Under the UCMJ, it is the human who is charged when things go sideways, not the AI. There is no Article 15 for the AI tool. If you submit a document generated by any AI tool that includes false info, violates OPSEC, or disobeys AI use policy, it’s your signature on that paper. That means you are liable.

EXAMPLE: An O-3 Company Commander under a tight deadline asks her XO to “use GenAI.mil to write a weekly SITREP.” The XO pulls up the tool, drops in a few bullet points, adds sensitive not CUI, classified info, and clicks submit. Now, that sensitive classified info is floating around in an IL5-contained dataset that’s still under audit.

A month later, during a routine AI inspection, the misstep is flagged. Here are some of the potential UCMJ violations for AI misuse and their reasons:

Article 92 – Disobeying lawful general orders

The XO potentially violated Article 92 for disobeying lawful general orders (usually an AI use policy or DoD cyber policy).

Article 107 – Making false official statements

If the AI-generated content misrepresents your input, falsely reports operations, or hallucinates as AI will sometimes do, any resulting false official statements are your responsibility.

Article 134 – The General Article

Irresponsible use of AI tools may well constitute conduct prejudicial to good order and discipline.

Security Clearance Review

Most of the AI platforms are unsecure and save the data that users enter. If sensitive or classified information ends up in the wrong hands, you’re looking at administrative separation and the loss of your security clearance. Losing access eligibility could drastically limit and reshape your career—not just in the military, but any private sector or contractor position requiring security clearance.

Defending Yourself Against Charges for AI Misuse

GenAI, just like any other tool in the DoW’s arsenal, is governed by policy, guidance, and, ultimately, the Uniform Code of Military Justice (UCMJ). The fact is: there is no AI defense in the courtroom. You’re not going to stand before a military judge and say, “Sir, the robot did it.”

This article is for informational and educational purposes only and does not constitute legal advice or establish an attorney-client relationship. If you are under investigation or believe you may be, you should immediately consult with a qualified military defense attorney or your local Trial Defense Services (TDS). At Tully Rinckey PLLC, we’re closely watching how AI is impacting administrative actions, investigations, and security clearance reviews. If you have additional questions, our team of military and security clearance attorneys is available to assist you today. Please call 8885294543 to schedule a consultation, or schedule a consultation online.

Michelle Alvarado Salermo, Esq. brings over 14 years of high-speed service in the U.S. Army, federal government, and private legal sector to her role as Associate Attorney in Tully Rinckey PLLC’s Buffalo office. A former U.S. Army Sergeant First Class, combat veteran, and former federal Claims Examiner for the U.S. Department of Labor, Michelle provides mission-driven legal counsel in federal litigation, military law, federal employment, corporate law, and emerging cybersecurity and technology law.