Loading story…
The Pentagon is rapidly expanding its use of generative AI across the Department of Defense, with officials saying adoption of the GenAI.mil platform has surged from about 80,000 users in late 2025 to roughly 1.5 million by June 2026, out of a workforce of around 3.5 million. Defense officials said the platform is already being used for administrative work including evaluation reports, medal citations, counseling statements, and reports mandated by Congress, while the department has also signed agreements with major AI vendors to deploy tools on classified networks for lawful operational use.
OpenAI said ChatGPT will be added to GenAI.mil in early July for more than 3 million civilian and military personnel, with certifications for controlled unclassified information and Impact Level 5. The expansion has drawn scrutiny because AI-generated submissions to Congress could contain errors or mischaracterizations that weaken oversight of military spending, particularly as the Defense Department seeks a record $1.5 trillion budget for fiscal year 2027; officials also noted that more than 100,000 AI agents have already been built on the platform as demand grows for higher token capacity and more efficient model use.
Track how attackers are adapting to this technology.
12 events from the most recent confirmed update back to the earliest known activity.
OpenAI said it plans to make ChatGPT available through the Pentagon's GenAI.mil platform in early July for more than 3 million defense civilian and military personnel. The deployment is being coordinated with the Defense Department's Chief Digital and Artificial Intelligence Office.
Senior defense officials said in late April that more than 1.3 million users were regularly using GenAI.mil and had built more than 100,000 AI agents. The figures showed rapid expansion of the Pentagon's internal AI platform.
WIRED reported that internal records from Peter Thiel's secretive Dialog society were left accessible online, exposing identities of more than 200 registered participants and sensitive personal details. The leaked trove reportedly included documents tied to a private retreat and an associated app.
After targeting Berkadia, ShinyHunters later published data it claimed was stolen from the company's Salesforce instance. The leaked dataset allegedly contained more than 300,000 unique email addresses plus names, physical addresses, phone numbers, and other information.
An OSINT Team post described a reported BreachForums breach exposing a database of 323,986 users, attributed to a threat actor named 'James' and hosted on infrastructure tied to ShinyHunters. The article said Resecurity validated the database as authentic and discussed tooling built to analyze the leaked data.
In June 2026, Pentagon officials said use of commercial AI tools through GenAI.mil had grown to 1.5 million users out of a workforce of about 3.5 million. The figure reflected major growth from the 80,000 users reported for December 2025.
On May 1, the Department of Defense signed agreements with eight major AI companies to deploy AI tools on classified networks for lawful operational use. Ars Technica reported Anthropic was excluded after reportedly refusing unrestricted military uses of Claude models.
Salesforce warned customers that attackers were targeting misconfigured Experience Cloud sites through the /s/sfsites/aura endpoint. Salesforce and Mandiant said the activity resulted from customer misconfiguration, not an inherent platform vulnerability, and noted misuse of a modified AuraInspector tool.
In March 2026, commercial real estate finance company Berkadia was targeted in a ShinyHunters 'pay or leak' extortion campaign. The incident involved alleged theft of data from Berkadia's Salesforce instance rather than ransomware encryption.
Pentagon officials said use of commercial AI tools through GenAI.mil stood at 80,000 users in December 2025. This provides an early adoption benchmark for the Defense Department's internal generative AI rollout.
The Defense Department launched the GenAI.mil platform in December 2025 with plans to integrate Gemini for Government. The platform was introduced to provide generative AI capabilities to defense personnel.
ShinyHunters claimed it started abusing insecure Salesforce Experience Cloud configurations via the Aura endpoint in September 2025. The actor said this marked the beginning of an ongoing data-theft campaign against organizations using misconfigured guest access.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
Follow how adversaries are adapting to this technology, and where it touches your stack today.
8 references tracked. Mallory keeps watching after this page renders.
nextgov.com
Open sourcearstechnica.com
Open sourcewired.com
Open sourcehaveibeenpwned.com
Open sourceosintteam.blog
Open sourcegoogle.com
Open sourcecloud.google.com
Open sourcehelp.salesforce.com
Open sourceMap indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
The U.S. Department of Defense has added **ChatGPT Mil**, a dedicated OpenAI deployment, and xAI's **Grok for Government** to its GenAI.mil enterprise portal alongside Google Gemini. The platform is intended to provide more than 3 million military personnel, civilian employees, and contractors with approved generative-AI tools for unclassified and controlled-unclassified-information work, including planning, policy, logistics, administration, and document processing. ChatGPT Mil operates separately from OpenAI's commercial service; DISA is building its environment and the NSA is assessing intrusion and containment risks. ChatGPT Mil and Grok for Government reportedly received Impact Level 5 authorization, allowing their environments to store and process sensitive, non-public unclassified data. Pentagon and OpenAI officials said the GenAI.mil deployments are not intended for combat use because of operational and connectivity constraints, while the department continues evaluating AI for operational applications. The multi-vendor strategy is designed to provide distinct frontier-model capabilities without vendor lock-in; Anthropic's Claude was not included following a dispute over restrictions on mass surveillance and lethal autonomous weapons.
Aug 31, 2026
The US Department of Defense has officially launched GenAI.mil, a generative artificial intelligence platform designed for military use and powered by Google’s Gemini AI application. Secretary of Defense Pete Hegseth issued a memorandum instructing all military personnel, civilians, and contractors to immediately incorporate the platform into their daily workflows, emphasizing its role in maintaining a technological edge over adversaries. GenAI.mil is authorized to handle controlled unclassified information and is accessible from all Pentagon and global military base desktops, with plans to expand its capabilities by integrating additional advanced AI models in the future. The platform offers features such as natural language conversation, retrieval-augmented generation (RAG), and secure web access, while meeting stringent security requirements including CUI and IL5 certifications for operational environments. The Department of Defense is also rolling out free AI training programs for all staff to maximize adoption and operational effectiveness. This initiative is part of a broader strategy to transform the DoD into an "AI-first" organization, aligning with recent White House directives to secure US leadership in AI-driven defense capabilities.
Jun 29, 2026
OpenAI released **GPT-5.4**, positioning it as a capability and efficiency upgrade aimed at knowledge-work use cases, including improved visual understanding (supporting analysis of very large images) and a claimed reduction in factual errors versus prior versions. The model is being rolled out across *ChatGPT* (web and apps), *Codex*, and the API, with higher-tier variants (e.g., “Thinking” and “Pro”) available to paid and enterprise/education offerings; the coverage also notes competitive pressure from **Anthropic** and heightened scrutiny around U.S. government/military AI use. Separately, OpenAI is reportedly developing an internal (and potentially customer-facing) alternative to **GitHub** after repeated GitHub outages disrupted engineering workflows, with the reliability issues tied to GitHub’s ongoing migration from a legacy data center to **Microsoft Azure** and associated configuration/underlying cloud incidents. In parallel, U.S. policymakers are drafting stricter, worldwide export controls for advanced AI accelerators from U.S. vendors (including **Nvidia** and **AMD**), proposing a tiered licensing regime with increasing compliance obligations (e.g., transparency requirements and potential inspections) as deployment scale grows—measures that could materially affect large AI infrastructure buildouts by major operators.
Jun 29, 2026