The FDA rolled out a new AI tool—but this doc’s not buying the hype
Industry Buzz
"I am a fan of using AI in specific instances for medical research purposes. But there's a lot of work that goes into training that model and then trying to get the right data points out of it in a more efficient manner. And I don't know if this tool is really ready for that." — Zachary Rubin, MD, pediatric allergist and immunologist
“This is a definite risk, especially given that there is currently no regulation on this AI." — William Soliman, PhD, CEO at the Accreditation Council for Medical Affairs
The FDA’s latest tech rollout sounds like something out of a sci-fi movie. The agency recently launched ELSA, a generative AI tool designed to speed up clinical protocol reviews, summarize adverse events, conduct label comparisons, generate code, and even help inspectors prioritize high-yield targets.
According to the agency, ELSA is already slashing review times. One scientific reviewer reportedly cut a task from 3 days to just 6 minutes. And the FDA says it all happened ahead of schedule and under budget.
But not everyone’s on board.
Zachary Rubin, MD, a pediatric allergist and immunologist, took to Instagram to ask a pointed question: Was this the same AI tool that reportedly fabricated citations in the controversial MAHA Commission Report?
Dr. Rubin emphasized that he isn’t anti-AI, but it has to be done right. "I am a fan of using AI in specific instances for medical research purposes. But there's a lot of work that goes into training that model and then trying to get the right data points out of it in a more efficient manner. And I don't know if this tool is really ready for that," he said.
William Soliman, PhD, CEO at the Accreditation Council for Medical Affairs (ACMA), echoed a similar sentiment:
“This is a definite risk, especially given that there is currently no regulation on this AI ... We have seen in other fields, like law, for example, where attorneys have lost their privilege to practice because they used AI to prepare legal briefs without checking the accuracy of the information. In some cases, the AI was making up reference cases outright.” []
Dr. Rubin also noted that multiple sources within the FDA have expressed concerns about ELSA’s rollout and maturity—concerns that have not been addressed publicly.
"This just feels like a giant slap in the face to the researching community," wrote Instagram user @ryankyleon in a comment on the video.
What physicians should watch for
ELSA is already being used to review clinical protocols, assess adverse events, and aid safety evaluations. In theory, this could have downstream effects on how quickly new treatments move through the review pipeline—or how thoroughly they’re assessed.
If you're involved in clinical trials or regulatory strategy, the tool’s influence on FDA workflows may soon be impossible to ignore.
But as Dr. Rubin points out, the lack of transparency around the tool’s development and limitations should be a key concern for medical professionals who rely on rigorous, unbiased reviews.
Related: AI-induced psychosis is a growing threat. Here's how to recognize it in patients—and how to helpThe bottom line
While the FDA touts ELSA as a leap forward in regulatory efficiency, questions remain about the model’s reliability, security, and oversight.
Dr. Rubin’s call for accountability may signal the beginning of a larger conversation among clinicians, researchers, and regulators. “We’ll see what happens when time goes by,” Dr. Rubin said. “And if I hear anything else, I’ll let you know.”
Doctors who rely on FDA decisions may want to keep an eye on what this AI tool is really learning—and how it’s being used.
Related: AI is transforming drug discovery—here’s how