HHS rolls out ChatGPT across the department
The Department of Health and Human Services has made ChatGPT available to all of its employees effectively immediately, according to a Tuesday departmentwide email obtained by FedScoop.
In that message, HHS Deputy Secretary Jim O’Neill said the rollout of the generative AI platform follows a directive from President Donald Trump’s AI Action Plan for agencies to ensure that workers who could benefit from the technology have access to it.
“This tool can help us promote rigorous science, radical transparency, and robust good health,” O’Neill said. “As Secretary Kennedy said, ‘The AI revolution has arrived.’”
O’Neill provided workers with instructions on how to log on to use the tool, as well as some warnings about how to treat outputs. He told workers to “be skeptical of everything you read, watch for potential bias, and treat answers as suggestions,” and directed them to weigh original sources and counterarguments prior to making a major decision.
He also pointed to potential attributes of the chatbot. “Like other LLMs, ChatGPT is particularly good at summarizing long documents,” O’Neill said, adding that large language models have already been useful for agencies such as the Food and Drug Administration and the Administration for Children and Families.
Per O’Neill’s message, the agency has taken security protocols for the tool. HHS Chief Information Officer Clark Minor has ensured that work with the AI platform “is carried out in a high-security environment,” and ChatGPT was granted authority to operate, or ATO, at a FISMA moderate level after a review of OpenAI’s security controls, O’Neill said.
There are limitations with what information can be used with the platform, according to the email. ChatGPT isn’t currently approved for use with sensitive personally identifiable, classified, export-controlled, or confidential commercial information. And agencies subject to HIPAA rules may not disclose protected health information while using the tool, O’Neill said.
HHS did not immediately respond to a request for comment on the news. OpenAI directed FedScoop to a LinkedIn post on the news from its vice president of government, Joseph Larson.
Linking to FedScoop’s reporting, Larson said HHS was the first agency to take advantage of the company’s $1 ChatGPT Enterprise offering to the federal government announced last month.
“This matters. HHS staff sit at the center of science and medical research, wading through endless studies, filings, and data,” Larson wrote in the post. “A tool that can summarize, surface insights, and translate complex science into plain English can help speed up discovery and strengthen public health.”
The $1 deal provides access to the generative AI tool at that price per agency for one year. It was the first of several deeply discounted AI offerings announced in recent weeks via the General Services Administration’s OneGov Strategy — a bid to improve how the government buys technology by leveraging standardized terms and pricing.
Since OpenAI’s initial deal, other companies — such as Anthropic, Google and Microsoft — have similarly offered their services to the government at deep discounts for a limited time.
“The best way to make AI work for government is to let the public servants closest to the work test these tools, find the limits, and help set the rules for how they’re used,” Larson said.
HHS’s announcement is “just the start,” he said, and OpenAI is “excited to keep building more partnerships across the federal government.”
This story was updated Sept. 9 to add comments from OpenAI’s Joseph Larson.
