What the First FDA AI Warning Letter Means for Pharma

On 2 April 2026, the U.S. FDA issued its first, and to-date only, warning letter citing inappropriate use of artificial intelligence in pharmaceutical manufacturing, naming Purolea Cosmetics Lab.

Investigators found the firm had used AI agents to draft drug specifications, procedures, and production records without adequate human review, alongside separate CGMP failures.

For regulatory affairs professionals watching AI move into regulated workflows, the FDA AI warning letter offers the first concrete evidence of where regulators draw the accountability line.

Image

What the First and Only FDA AI Warning Letter Shows

The Food and Drug Administration (FDA) inspected Purolea Cosmetics Lab in Livonia, Michigan, from October 28 to 30, 2025.

The inspection identified multiple Current Good Manufacturing Practice (CGMP) deficiencies, including inadequate quality-unit oversight, insufficient testing, and failures involving manufacturing controls. The FDA also found insanitary conditions and issues with unapproved drug products.

The AI finding was specific. Purolea told investigators that it had used AI agents to create drug specifications, procedures, and master production or control records intended to meet FDA requirements.

The agency, however, stated that AI-generated documents still require appropriate human review. The agency linked the failure to review those outputs to 21 CFR 211.22(c), which concerns the responsibilities of the quality control unit.

The inspection also found that process validation had not been completed before drug distribution. According to the FDA, Purolea said its AI system had not informed it that process validation was required.

So while the warning letter does not establish a new rule that pharmaceutical manufacturers cannot use AI, it establishes that AI assistance does not replace regulated human accountability.

View interview How Should Pharma Companies Audit AI Vendors?

How Should Pharma Companies Audit AI Vendors?

“This is one of the riskiest industries there is because you're dealing with human lives, you're dealing with experimental protocols, and you're dealing with regulatory bodies where you might not get another shot at that clinical trial."

Image

AI Cannot Operate as the Quality Unit

The FDA AI warning letter has significance beyond document drafting.

CGMP systems depend on controlled procedures, appropriate review, documented decisions, and qualified personnel.

Those requirements apply whether information is produced manually, through conventional software, or with generative AI.

The FDA therefore instructed that, if Purolea resumes manufacturing and uses AI for CGMP activities, AI outputs or recommendations must be reviewed and cleared by an authorised human representative of the firm's quality unit. 

This aligns with the agency’s AI policy. Its 10 guiding principles for good AI practice emphasise human-centric design, risk-based approaches, data governance, documentation, performance assessment, lifecycle management, and clear information about AI systems. 

For manufacturers, the practical implications include:

• Define what the AI system is permitted to do.
• Establish qualified human review before regulated use.
• Maintain traceable records of AI-generated content and subsequent changes.
• Assess whether AI outputs are accurate, complete, and appropriate for their intended use.
• Keep existing quality-unit responsibilities clearly assigned.

The Regulatory Signal for Pharmaceutical Manufacturing

The warning letter arrives as AI moves deeper into manufacturing, quality systems, process monitoring, and Technical Operations.

The FDA has previously discussed the application of a risk-based regulatory framework to AI in drug manufacturing, including questions around existing CGMP requirements and the governance of AI-enabled systems.

The Purolea case provides a concrete compliance example. The regulatory question is not simply whether AI can produce acceptable text but whether the pharmaceutical quality system can demonstrate that AI-assisted work remains controlled, reviewed, traceable, and fit for purpose.

This also reinforces the importance of data integrity and controlled documentation as manufacturing becomes more digital. Pharmatica has previously examined how AI, digital twins, connected equipment, and other technologies are increasing the complexity of manufacturing data and auditability.

The wider message for the industry is that AI can support GMP operations, but it does not own the decision. Responsibility remains with the regulated organisation and its qualified personnel.

What the First AI-related CGMP Warning Means for Industry

The Purolea warning letter does not create a standalone AI regulation for pharmaceutical manufacturing. It applies existing CGMP expectations to an AI-enabled workflow.

That makes the event significant. It demonstrates that regulators can address problematic AI use through established quality requirements, even while AI-specific manufacturing frameworks continue to develop.

The immediate priority is therefore not simply selecting compliant AI software but establishing governance around how AI enters regulated workflows, who reviews its outputs, what evidence supports its use, and how those decisions are documented.

At Pharmatica, we track the regulatory, technological, and operational changes reshaping pharmaceutical manufacturing. Our Insights connect emerging AI capabilities with the quality systems and regulatory responsibilities that determine whether innovation can operate safely at scale.

Pharmatica: Insight. Connection. Impact.