Homeland Security cuts off access to ChatGPT and other commercial AI 

Staff at the Department of Homeland Security are no longer allowed to use commercial generative artificial intelligence tools like ChatGPT and Claude, according to a memo sent to employees this month. 

The move is a reversal of a previous policy — which had conditionally allowed the use of commercial systems — and a pivot toward technology developed in-house. 

Earlier this month, DHS’s chief information officer, Antoine McCord, sent a memo directing component tech offices to begin “restricting” the use of generative AI systems and pointing employees to internal tools. Older guidance, which the CIO described as “outdated” and “too narrowly” focused on commercial generative AI, was also removed from an online list of IT management directives. 

The decision comes as federal agencies weigh various pathways toward integrating generative AI into their workflows, a priority of both the Biden and Trump administrations. While some government agencies initially blocked generative AI systems, CIOs have slowly started to develop usage policies. Some agencies, like DHS and the General Services Administration, have now built their own platforms based on commercial technologies, while others have opted to use products like ChatGPT Gov through government cloud systems. 

Late last year, Michael Boyce, the leader of DHS’s AI Corps, announced that a new tool called DHSChat had become available to the agency’s nearly 20,000 employees. The platform was designed with commercial large language model technology through an API and built in a secure environment, FedScoop reported at the time. Its introduction came after DHS conducted several generative AI pilots throughout its component agencies. 

Now, in a pivot to focus more on DHSChat, the agency has revoked previous approvals for commercial generative AI systems, the latest memo explained. “Commercial GenAI tools are defined as GenAI technology or products available for use or purchase by the general public,” said the document, which also clarified that AI systems created as part of customized software or developed specifically for use by the government do not fall into this category. 

Staff were told to cancel payments for existing commercial generative AI systems and to remove training courses.

DHS is now in the process of developing a new generative AI policy in accordance with the Trump administration’s recent White House guidance on accelerating the federal government’s use of the technology. 

This story was supported by the Tarbell Center for AI Journalism.

Madison Alder contributed reporting.