FDA fast-tracks AI tools to slash regulatory scientific review time

The U.S. Food and Drug Administration (FDA) is moving quickly to bring AI tools to every center in the agency following the success of its first AI-assisted scientific review pilot. By June 30 this year, the FDA aims to have all centers using a shared, secure generative AI system designed to speed up regulatory reviews without compromising scientific rigor.

“I was blown away by the success of our first AI-assisted scientific review pilot,” said FDA Commissioner Dr. Martin A. Makary in a press release. “We need to value our scientists’ time and reduce the amount of non-productive busywork that has historically consumed much of the review process.”

Days to minutes, a scientific review process transformed

In the pilot, generative AI tools were used to handle some of the more labor-intensive parts of the scientific review process. Jinzhong (Jin) Liu, Deputy Director at the Center for Drug Evaluation and Research, called the results “a game-changer,” noting that generative AI “has enabled me to perform scientific review tasks in minutes that used to take three days.”

This kind of time savings could bring new treatments to patients more quickly, a vital improvement given the stacks of AI-related drug applications piling up for the FDA.

A unified platform and a familiar face

Makary has expressed frustration on X.com over how long it takes for new therapies to be approved and emphasizes that when it comes to AI integration, time is of the essence.

“There have been years of talk about AI capabilities in frameworks, conferences and panels but we cannot afford to keep talking,” Dr. Makary said. “It is time to take action.”

That action includes launching a unified AI platform across the agency. The system will integrate directly with FDA’s internal data and offer features tailored to each center’s needs, from document processing to content summarization. Usability improvements, enhanced document integration, and tighter security controls are also planned.

The FDA did not announce what tech company will support their platform, but Wired reported last week that the FDA is in talks with OpenAI, parent company of ChatGPT. This partnership coincides with ChatGPT’s release of Deep Research, a new feature that conducts real-time web research and returns answers with citations.

Leadership is not waiting around

To help drive this fast-track rollout, Makary is working alongside Jeremy Walsh, the newly appointed Chief AI Officer and former Chief Technologist at Booz Allen Hamilton, and Sridhar Mantha, who formerly led the Office of Business Informatics in the Center for Drug Evaluation and Research.

The rollout strategy focuses not only on speed but also on adaptability. The FDA plans to collect staff feedback and refine tools as they scale, allowing the system to evolve and better meet the practical realities of scientific review.

Setting the tone for federal agencies

The FDA will share more updates on the rollout in June, as integration continues across all centers. At the same time, additional use cases for generative AI are already under consideration, building on earlier tools like the Computerized Labeling Assessment Tool, which uses machine learning to spot medication labeling errors.

If the results of the FDA’s scientific review pilot are positive and its bet on AI helping it keep pace with tech innovation pays off, the impact may reach well beyond the agency. The FDA revolutionizing its ability to manage its workload could serve as a model for other federal agencies on how to approach their own AI integration.