Artificial Intelligence is rapidly becoming part of the medical device and pharmaceutical industry's daily operations. Companies are using AI to draft procedures, generate documentation, analyze data, develop software, and even assist with manufacturing activities.
But the FDA's first warning letter specifically addressing the misuse of AI in a regulated manufacturing environment delivers an unmistakable message:
AI can support your quality system—but it cannot replace it.
What Happened?
In April 2026, the FDA issued a Warning Letter to Purolea Cosmetics Lab after investigators determined that the company relied on AI-generated pharmaceutical manufacturing documentation without appropriate human review or Quality Unit oversight.
According to the FDA, AI had been used to generate:
- Drug specifications
- Manufacturing procedures
- Master production records
- Other GMP documentation
When inspectors questioned deficiencies in the documentation, the company reportedly responded that the AI system had not identified the applicable regulatory requirements.
The FDA rejected that explanation entirely.
The FDA Isn't Regulating AI…
One of the biggest misconceptions surrounding this warning letter is that the FDA has introduced new AI regulations.
It hasn't.
Instead, the Agency simply applied existing Current Good Manufacturing Practice (cGMP) requirements.
Whether documentation is written by an employee, a consultant, or an AI platform, the manufacturer remains fully responsible for ensuring that it is:
- Accurate
- Complete
- Scientifically sound
- Approved through the Quality System
- Maintained under document control
The regulatory obligation never shifts from the manufacturer to the technology.
Human Oversight Is Still Required
Generative AI can dramatically improve efficiency.
It can:
- Draft SOPs
- Generate validation protocols
- Create software documentation
- Assist with risk analyses
- Summarize technical reports
However, every AI-generated output must be treated as a draft—not as an approved regulatory document.
Organizations should establish procedures requiring qualified personnel to:
- Verify technical accuracy
- Confirm regulatory compliance
- Validate references and standards
- Review for completeness
- Approve documents through normal document control processes
AI may write the first draft.
Humans remain responsible for the final product.
This Applies Beyond Pharmaceutical Manufacturing
Although this Warning Letter involved cGMP documentation for drug manufacturing, the underlying principle extends across regulated industries.
Medical device manufacturers using AI to generate:
- Design History File (DHF) documentation
- Design Inputs and Outputs
- Software Requirements Specifications (SRS)
- Verification and Validation documentation
- Risk Management Files
- Clinical Evaluation Reports
- Technical Documentation
Remain fully accountable for the accuracy, completeness, and regulatory compliance of every document submitted to regulators or maintained within the Quality Management System.
The same expectations apply under FDA Quality System requirements, ISO 13485, IEC 62304, ISO 14971, and other applicable standards.
AI Governance Is Becoming Part of Quality Management
Rather than banning AI, the FDA is reinforcing a more practical expectation:
Organizations must govern AI just like any other quality-affecting process.
That means companies should establish policies addressing:
- Approved AI tools
- Acceptable use cases
- Human review requirements
- Documentation verification
- Version control
- Data security and confidentiality
- Employee training
- Record retention
- Change management
As AI becomes more integrated into regulated workflows, these controls will become an increasingly important component of Quality Management Systems.
The Takeaway
The FDA's first AI-focused Warning Letter isn't really about artificial intelligence.
It's about accountability.
Using AI does not reduce a manufacturer's regulatory obligations. If anything, it increases the need for documented oversight, robust review processes, and effective Quality System controls.
AI can accelerate document development and improve productivity—but regulatory compliance still depends on qualified professionals applying engineering judgment, scientific review, and independent quality oversight.
Companies that implement AI within a controlled, risk-based Quality Management System will be well positioned to benefit from the technology while remaining compliant.
Those that rely on AI without adequate governance may find that the FDA views AI-generated mistakes no differently than human ones.
At RQMIS, we help medical device, IVD, pharmaceutical, and biotechnology companies implement AI-enabled workflows that remain fully aligned with FDA, ISO, and global regulatory requirements. If your organization is incorporating AI into product development or quality processes, now is the time to ensure your Quality Management System evolves with it.