NEW YORK (AP) — The US Department of Health and Human Services on Thursday outlined a strategy expand its use artificial intelligencefollowing the Trump administration enthusiastic embrace rapidly evolving technology while raising questions about how health information would be protected.
HHS called the plan a “first step” aimed primarily at streamlining its work and coordinating the adoption of AI across divisions. But the 20-page document also teased some larger plans to boost AI innovation, including analyzing patient health data and drug development.
“For too long, our department has been overwhelmed by bureaucracy and busywork,” HHS Deputy Secretary Jim O’Neill wrote in an introduction to the strategy. “It’s time to break down these barriers to progress and unite in our use of technology to make America healthy again.”
The new strategy signals how leaders across the Trump administration have embraced AI innovation, encouraging employees across the federal workforce to use chatbots and AI assistants for their day-to-day tasks. As generative artificial intelligence technology has made significant leaps under President Joe Biden, he issued an executive order to create guardrails for their use. But when President Donald Trump took office, he canceled that order and his administration has sought to remove barriers to the use of AI across the federal government.
Experts said the administration’s willingness to modernize government operations presents both opportunities and risks. Some said artificial intelligence innovations within HHS required strict standards because they dealt with sensitive data, and doubted whether these would be met under Health Secretary Robert F. Kennedy Jr. Some in Kennedy’s own “Make America Health Again” movement have also expressed concern about tech companies’ access to people’s personal information.
HHS’s new plan calls for adopting a “try-first” culture that will help employees become more productive and capable through the use of AI. Earlier this year, HHS made the popular AI model ChatGPT available to all employees in the department.
The document identifies five key pillars for its future AI strategy, including creating a governance structure that manages risk, designing an AI resource set for use across the department, empowering staff to use AI tools, funding programs to set standards for the use of AI in research and development, and embedding AI in public health and patient care.
HHS divisions are already said to be working to promote the use of artificial intelligence “to provide personalized, context-aware health guidance to patients through secure access and real-time interpretation of their medical records.” Some in Kennedy’s Make America Healthy Again movement have raised concerns about using artificial intelligence tools to analyze health data and say they are uncomfortable with the U.S. Department of Health working with big tech companies to access people’s personal information.
HHS has previously faced criticism for crossing legal lines when sharing sensitive data handed over the personal health information of Medicaid recipients immigration and customs officials.
Oren Etzioni, an artificial intelligence expert who founded a nonprofit to combat political deepfakes, said HHS’s enthusiasm for using AI in healthcare is worth celebrating, but cautioned that speed should not come at the expense of security.
“The HHS strategy sets ambitious goals—a centralized data infrastructure, rapid deployment of AI tools, and an AI-enabled workforce—but ambition brings risk when dealing with the most sensitive data Americans have: their health information,” he said.
Etzioni said the strategy’s call for “gold standard science”, risk assessment and transparency in AI development appear to be positive signs. But he said he doubts HHS can meet those standards under Kennedy, who he says often flouts rigor and scientific principles.
Darrell West, a senior fellow at the Brooking Institution’s Center for Technological Innovation, noted that the document promises to strengthen risk management but does not detail how it will be done.
“There are many unanswered questions about how sensitive medical information will be handled and how the data will be shared,” he said. “There are clear safeguards for individual records, but not as many protections for aggregated information analyzed by AI tools. I’d like to understand how officials plan to balance the use of medical information to improve operations with privacy protections that protect people’s personal information.”
But West said that if done carefully, “it could become a transformative example of a modernized agency that is operating at a much higher level than before.”
The strategy states that HHS had 271 active or planned AI implementations in fiscal year 2024, a number projected to grow by 70% in 2025.
