FDA balances faster AI adoption with safety as healthcare technology advances
The Trump administration is pushing to accelerate artificial intelligence (AI) adoption across healthcare, but the effort comes as federal agencies face staffing attrition and questions about their capacity to oversee increasingly sophisticated medical technologies.
Health Exec spoke with Aubrey Schick, a digital health policy expert in regulatory and product strategy and former senior digital health advisor at the U.S. Food and Drug Administration (FDA), who offered insights into how FDA has changed under the Trump administration, but is evolving to take on the growing regulatory need for AI. She said the agency retains substantial expertise in regulating AI-enabled medical devices. However, continued staff attrition and the rapid evolution of generative AI and autonomous healthcare systems present ongoing challenges for regulators.
Schick previously worked at the FDA’s Digital Health Center of Excellence, which supports the agency’s oversight of digital health technologies. Although she said no one from that center departed under the Department of Government Efficiency (DOGE) initiatives, the team has experienced attrition over the past two years of the Trump administration. But, she added the FDA has the staff to handle the increasing numbers of AI submissions.
“There’s a lot of expertise at FDA in terms of how AI-enabled medical devices are regulated,” Schick said.
She noted that the agency has been evaluating AI-enabled medical technologies since the 1990s and numbers of submissions have been steadily increasing over the past decade. The FDA continues to maintain its publicly available list of AI-enabled medical devices, providing insight into the technology’s expanding clinical footprint. Meanwhile, the agency is developing its approach to newer applications, including generative AI.
Risk-based regulation remains central with AI
As the number and variety of AI-enabled medical devices increase, the FDA evaluates products according to their intended use, clinical indication and potential risks, rather than applying a single standard to every algorithm.
Schick said AI-enabled software is reviewed through the same general regulatory framework used for traditional medical devices. The evidence requirements and approval pathway depend on the technology and its intended application.
Depending on the product, manufacturers may pursue a 510(k) clearance based on substantial equivalence to an already existing FDA cleared product, a de novo classification for certain novel devices, or premarket approval (PMA) for higher-risk products.
The FDA also considers what special controls and post-market monitoring may be necessary to identify potential problems after a technology reaches clinical practice.
Schick said the agency has continued to take a risk-based approach, allowing additional scrutiny for products with greater potential to affect patient safety.
As healthcare organizations seek faster access to AI tools, maintaining that level of review will be important, particularly as systems move beyond conventional image analysis into more complex clinical decision-making and autonomous procedures.
ARPA-H programs could inform future pathways
The Advanced Research Projects Agency for Health (ARPA-H) is also supporting efforts to advance AI applications in healthcare, including its ADVOCATE program, which is assessing new approaches to AI-enabled cardiology healthcare delivery and supervision.
Schick described the initiative as an opportunity to generate evidence and develop frameworks that could inform future FDA regulatory pathways.
However, she emphasized that the program itself is not a regulatory pathway. Instead, findings about how AI agents are supervised and the environments in which they operate could help shape future regulatory approaches.
She also pointed to the potential importance of expanded FDA authority for post-market oversight as AI systems become more complex and continue to evolve after deployment.
Generative AI and autonomous systems pose new questions
The next wave of healthcare AI includes large language models, autonomous imaging technologies and systems designed to assist with procedures such as mechanical thrombectomy for stroke patients. These technologies could eventually help extend specialized care to rural and underserved areas where experienced clinicians or advanced procedural services may not be readily available.
But determining the appropriate evidence requirements, level of human supervision and post-market monitoring for such applications remains an evolving process. Schick said the FDA has been seeking input from industry, researchers and other experts as it develops its thinking on generative AI. She added that manufacturers often engage with the agency through pre-submission discussions, allowing regulators to consider potential evidence requirements and oversight needs before formal applications are submitted.
Despite the pressure to accelerate adoption, Schick said the FDA’s review teams remain focused on public health and bringing beneficial technologies to patients safely and effectively.
For developers, healthcare providers and imaging specialists, the challenge will be to advance AI innovation while ensuring that regulatory oversight keeps pace with the technology’s growing capabilities.
Dave Fornell has covered healthcare for more than 17 years, with a focus in cardiology and radiology. Fornell is a 5-time winner of a Jesse H. Neal Award, the most prestigious editorial honors in the field of specialized journalism. The wins included best technical content, best use of social media and best COVID-19 coverage. Fornell was also a three-time Neal finalist for best range of work by a single author. He produces more than 100 editorial videos each year, most of them interviews with key opinion leaders in medicine. He also writes technical articles, covers key trends, conducts video hospital site visits, and is very involved with social media. E-mail: [email protected]