Welcome to Health Bytes. I’m your host, and today we’re talking about Elsa, the FDA’s new agency-wide AI tool that’s stirring up controversy. Launched under Commissioner Marty Makary’s “aggressive timeline” to scale AI by June 30, Elsa is billed as a game-changer for scientific reviewers and investigators. According to FDA Chief AI Officer Jeremy Walsh, “Today marks the dawn of the AI era at the FDA,” with Elsa set to accelerate clinical protocol reviews, summarize adverse events, and even generate code for nonclinical databases.

But inside the agency, staffers call the rollout “rushed,” “buggy,” and “overhyped.” One employee warned, “Makary and DOGE think AI can replace staff and cut review times, but it decidedly cannot.” Built on Anthropic’s Claude model and developed by Deloitte, Elsa still provides inaccurate summaries and lacks proper guardrails. As FDA centers grapple with glitches and disconnected systems, the promise of faster, safer reviews feels premature. That’s Health Bytes—stay tuned for more.
Link to Article