Kennedy at Senate hearing. From hearing livestream.
Health Secretary Robert F. Kennedy Jr. has moved quickly to reshape HHS and its agencies. He fired the CDC director and reconstituted the vaccine advisory panel with vaccine skeptics. The latest? He has turned on a department-wide ChatGPT service. The moves land alongside CMS’s AI-assisted prior-authorization pilot (with human review) and an White House-backed AI Action Plan and MAHA roadmap that signals a reassessment of childhood vaccine protocols.
Employees reportedly heard of the ChatGPT news in an email with the subject line “AI Deployment,” which told workers that ChatGPT would be rolled out for all employees at the agency. The deployment falls under Clark Minor, a former Palantir employee now serving as HHS chief information officer. The message, sent by Deputy Secretary Jim O’Neill, stated that AI is starting to improve health care, business and government operations. It highlighted how administrative tasks like emails and meetings can hinder productivity, urging staff to push back against obstacles to “making America healthy again.”
An ‘AI revolution has arrived’
The email touted ChatGPT as a tool to promote “rigorous science, radical transparency, and robust good health,” echoing Kennedy’s stance that the “AI revolution has arrived.” Access is straightforward: Staff can log in at go.hhs.gov/chatgpt using their government email, pose questions, and refine responses. The guidance stressed skepticism—treat outputs as suggestions, check for bias, and verify with original sources before big decisions. It noted the tool excels at summarizing long documents.
Security measures were outlined: Minor has set up a high-security environment for most internal data, including procurement-sensitive info and routine non-sensitive personally identifiable information. But it’s off-limits for sensitive PII like Social Security numbers, classified data, export-controlled material or trade secrets. HHS didn’t respond to requests for comment on the rollout.
The moves land alongside CMS’s AI-assisted prior-authorization pilot, known as the Wasteful and Inappropriate Service Reduction (WISeR) Model, launched June 27 for six years in six states, including Washington. It requires prior approval for 17 medical services flagged for overuse, using AI to screen requests with human review, per CMS FAQs. Critics warn it could deny care and echo biased systems from past trials. And there’s the White House-backed Make America Healthy Again (MAHA) roadmap, released Tuesday, which signals a reassessment of childhood vaccine protocols. The report calls for a new vaccine framework, reviewing the schedule, probing injuries and ensuring transparency, while targeting ultra-processed foods and pesticides.
Broader AI adoption at the federal level
This push aligns with broader federal AI adoption that started under Biden but has accelerated under Trump, including OpenAI’s $1-per-agency deal for ChatGPT Enterprise through next year. It’s noteworthy at HHS, where Kennedy has questioned established science, eased vaccine access barriers and promoted theories on autism causes. The agency also plans AI use at CMS to assess treatment eligibility.
According to an agency website, HHS aims to “leverage artificial intelligence (AI) capabilities to solve mission challenges and gain new insights into complex problems while ensuring that our solutions are ethical, effective, and secure.”
