> ## Content Index
> Fetch the complete content index at: https://www.fdaweb.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# AI Cited in Just 7 Warning Letters Since 2022: Study
- URL: https://www.fdaweb.com/ai-cited-in-just-7-warning-letters-since-2022-study/
- Published: 2026-11-04T23:14:47.000Z
- Updated: 2026-11-04T23:14:47.000Z
- Author: David McFarland
- Tags: enforcement

FDA referenced artificial intelligence in just seven of 3,011 Warning Letters issued between 1/2022 and 9/28/2026, with only one letter explicitly identifying a company's use of AI as a regulatory violation, according to a [new study](https://link.springer.com/article/10.1007/s44395-026-00061-5?ref=fdaweb.com) published in *Discover Pharmaceutical Sciences*.

The [retrospective analysis](https://link.springer.com/article/10.1007/s44395-026-00061-5?ref=fdaweb.com), published 10/31, found AI references in 0.23% of the letters: three involving medical devices, three involving tobacco products and one involving drugs.

The only letter directly citing improper AI use was issued in April to Purolea Cosmetics Lab ([see earlier story](https://www.fdaweb.com/multiple-issues-in-purolea-cosmetics-lab-inspection/)). FDA cited the company for using AI agents to generate specifications, procedures and master records without quality unit review and approval, a violation of 21 CFR 211.22(c).

FDA stated that AI-generated material used in current good manufacturing practice (CGMP) activities must be reviewed and cleared by an authorized human representative of the quality unit.

In the three device warning letters, AI references concerned diagnostic claims or unvalidated software capabilities. The tobacco letters identified AI chat functions among product features potentially appealing to young people, but did not cite AI use itself as a violation.

The researchers also identified two Form FDA 483s and one establishment inspection report referencing AI, although none cited AI use as a deficiency.

The study concluded that FDA has generally addressed AI-related compliance issues through existing regulatory requirements rather than AI-specific enforcement standards. “FDA references to AI in warning letters are rare, recent, and heterogeneous," the authors wrote.

The researchers cautioned that their analysis did not examine FDA's treatment of AI in product approvals or software as a medical device. They also noted that the available inspection records were not comprehensive.

The findings suggest that companies incorporating generative AI into regulated operations remain subject to established quality system requirements, including human review and approval of AI-generated CGMP documentation.