US FDA launches AI tool to reduce time taken for scientific reviews
Source: Reuters
The U.S. Food and Drug Administration said on Monday that it had launched a generative AI tool, Elsa, aimed at improving efficiency across its operations, including scientific reviews.
"Today's rollout of Elsa is ahead of schedule and under budget, thanks to the collaboration of our in-house experts across the centers," said FDA Commissioner Marty Makary.
The agency said it is already using Elsa to expedite clinical protocol reviews, shorten the time needed for scientific evaluations, and pinpoint high-priority inspection targets.
-snip-
Elsa assists with reading, writing, and summarizing tasks. It can summarize adverse events to support safety profile assessments of drugs and rapidly compare packaging inserts.
-snip-
Read more: https://www.reuters.com/business/healthcare-pharmaceuticals/us-fda-launches-ai-tool-reduce-time-taken-scientific-reviews-2025-06-02/
Dear God.
Generative AI models hallucinate, and it can be hard to catch the errors.
Just seconds before I saw that Reuters story in my Bluesky feed, I saw this post from science fiction writer John Scalzi commenting on an error by Google's AI overview that someone told him about:
WE DON'T TALK ABOUT BRIAN
— John Scalzi (@scalzi.com) 2025-06-02T22:09:13.229Z
not after "the incident"
I still have pudding in my ear canal
Also, yes, don't trust "AI" to tell you anything about anything, y'all
The person who told him about the error commented, "Seriously, getting something so basic wrong makes all the information useless, because I can't trust any of it."
And John.agreed:
Scalzi's post and the Reuters story were posted on Bluesky just 4 minutes apart.
US FDA launches AI tool to reduce time taken for scientific reviews reut.rs/4557Hx3
— Reuters (@reuters.com) 2025-06-02T22:05:10Z
= new reply since forum marked as read
