The Food and Drug Administration (FDA) launched the first agency-wide generative artificial intelligence (GenAI) tool on Monday to streamline efficiency for employees ranging from scientific reviewers to investigators.
FDA Commissioner Marty Makary said the agency launched the GenAI tool, known as Elsa, following a successful pilot program. The FDA said the tool’s rollout came nearly a month ahead of the FDA’s June 30 target to ramp up agency-wide AI use, and it came in under budget.
Hosted in a high-security GovCloud environment, the large language model tool enables FDA employees to access internal documents while preventing researchers from accessing private data submitted by industry members.
“Today marks the dawn of the AI era at the FDA with the release of Elsa, AI is no longer a distant promise but a dynamic force enhancing and optimizing the performance and potential of every employee,” said FDA Chief AI Officer Jeremy Walsh in a statement on June 2 announcing the tool’s launch.
“As we learn how employees are using the tool, our development team will be able to add capabilities and grow with the needs of employees and the agency,” he continued.
The FDA said the tool can summarize information to support safety profile assessments, perform fast label comparisons, and generate database code for nonclinical applications.
Those capabilities have enabled the agency to use Elsa to accelerate clinical protocol reviews, shorten scientific evaluations, and identify high-priority inspection targets, according to the FDA.
The FDA announced its plans to scale its use of AI across all 11 of its centers early last month, after announcing that Elsa had been successfully piloted. Makary had said at the time that the tool would later have expanded use cases and better functionality adapted to “the evolving needs of each center” after the June 30 timeline.
Beyond that, the agency said it will expand generative AI capabilities on a single unified platform, and that future updates will improve usability and expand document integration.