FDA Completes First AI-Assisted Review Pilot, Accelerates Rollout Across All Centers by June 2025

• By John K. Waters
• 05/09/2025

The U.S. Food and Drug Administration is moving fast to institutionalize generative AI across its operations, following the successful completion of its first AI-powered scientific review pilot, the agency announced in a press release. FDA Commissioner Martin A. Makary, M.D., M.P.H., announced that all agency centers are expected to deploy secure generative AI systems by June 30, 2025.

"I was blown away by the success of our first AI-assisted scientific review pilot," said Makary. "We need to value our scientists’ time and reduce the amount of non-productive busywork that has historically consumed much of the review process."

The pilot leveraged generative AI to accelerate scientific review tasks by dramatically cutting down the time required for common workflows. According to Dr. Jinzhong Liu, Deputy Director in the Center for Drug Evaluation and Research, the new system reduced multi-day tasks to a matter of minutes. “This is game-changing technology that has enabled me to perform scientific review tasks in minutes that used to take three days,” Liu said.

The rollout is being orchestrated by the FDA’s newly appointed Chief AI Officer Jeremy Walsh and Sridhar Mantha, who formerly led tech transformation at CDER. The deployment will standardize AI usage across centers using a unified, secure infrastructure integrated with FDA’s internal data platforms.

Makary made it clear that the agency is moving past theoretical discussions. "There have been years of talk about AI capabilities in frameworks, conferences and panels," he said.
"We cannot afford to keep talking. It is time to take action."

While the goal is full operational capability by the end of June, the FDA plans to continue refining the platform post-deployment—expanding document integration, improving prompt usability, and adapting model outputs to meet the specific needs of each center. The AI infrastructure will comply with strict internal security requirements and FDA policy guidelines.

Further technical details, performance metrics, and platform updates are expected to be released publicly in June.