See every side of every news story
Published loading...Updated

FDA’s artificial intelligence is supposed to revolutionize drug approvals. It’s making up studies

UNITED STATES, JUL 23 – The FDA's Elsa AI tool has produced fabricated studies and misrepresented research, raising concerns about drug approval reliability amid efforts to speed up the process, according to multiple FDA employees.

  • On June 2, 2025, FDA Commissioner Marty Makary deployed Elsa, an AI tool, across the agency to speed drug and device approvals.
  • Following a strong endorsement from the head of the Department of Health and Human Services to implement generative AI, this deployment proceeded despite ongoing doubts about the technology's accuracy and regulatory oversight.
  • Several current and former FDA officials have expressed concerns that Elsa commonly fabricates false studies and inaccurately represents research, limiting its reliability for important regulatory responsibilities.
  • Makary acknowledged Elsa's potential to hallucinate but said such errors may be reduced with more detailed queries, while Kennedy testified AI already speeds drug approvals.
  • These flaws raise risks of unsafe drug approvals, prompting calls for stricter AI regulation and systematic postmarket evaluations to protect public health.
Insights by Ground AI
Does this summary seem wrong?

11 Articles

Left

An investigation by CNN reveals the slipping of Elsa, a new AI tool that was supposed to boost the effectiveness of the Food and Drug Administration.

·Paris, France
Read Full Article
Think freely.Subscribe and get full access to Ground NewsSubscriptions start at $9.99/yearSubscribe

Bias Distribution

  • 100% of the sources lean Left
100% Left

Factuality 

To view factuality data please Upgrade to Premium

Ownership

To view ownership data please Upgrade to Vantage

CNN broke the news in Atlanta, United States on Wednesday, July 23, 2025.
Sources are mostly out of (0)

You have read 1 out of your 5 free daily articles.

Join millions of well-informed readers who use Ground to compare coverage, check their news blindspots, and challenge their worldview.