How often do pharmaceutical regulators actually identify artificial intelligence as a GMP compliance problem? A recent analysis of 3,011 FDA Warning Letters found only one direct CGMP violation involving AI. But evidence from the UK, Switzerland, Japan, Europe and other regulatory authorities reveals a more complex picture: inspectors are already encountering AI-related failures, while regulatory expectations are developing faster than publicly reported enforcement cases.
On 31 October 2026, a peer-reviewed study published in Discover Pharmaceutical Sciences examined references to artificial intelligence in FDA regulatory correspondence.
The researchers reviewed 3,011 Warning Letters issued between 1 January 2022 and 28 September 2026.
Only seven letters mentioned AI, and just one directly identified a company’s use of AI as a CGMP violation.
These findings raise an important question:
Are pharmaceutical regulators in other countries identifying similar compliance problems, or is AI still largely absent from publicly available GMP inspection findings?
AIforPharma.org reviewed publicly available regulatory publications, inspection-related communications, enforcement examples and GMP policy documents from 12 regulatory authorities and networks.
The results show that regulators are addressing AI in substantially different ways.
Some have already documented actual compliance problems. Others are developing explicit expectations for AI in manufacturing, requiring notification of significant changes, or introducing AI into their own regulatory oversight systems.
However, published information does not yet support a reliable worldwide estimate of AI-related GMP deficiencies.
United States: one direct CGMP case in 3,011 FDA Warning Letters
The retrospective study by Kwiecinski and Nandy provides an unusually useful quantitative baseline.
Of 3,011 FDA Warning Letters examined, seven referred to AI, representing approximately 0.23% of the dataset.
The seven letters comprised:
- Three medical-device Warning Letters.
- Three tobacco-product Warning Letters.
- One drug-related Warning Letter.
Importantly, the references to AI did not all represent violations involving the inappropriate use of AI.
In the medical-device cases, AI-related product claims were relevant to regulatory classification or authorisation.
In the tobacco cases, AI-related features were discussed in connection with unauthorised products.
Only one drug Warning Letter directly addressed AI used in CGMP activities.
The Purolea Cosmetics Lab case - 2 April 2026
FDA identified the company’s use of AI agents to generate drug product specifications, procedures and master production or control records.
The agency found that the AI-generated documents had not been subjected to the required Quality Unit review and approval.
FDA explicitly cited 21 CFR 211.22(c).
The violation was not the use of AI itself, but the failure to ensure that AI-generated GMP documents were appropriately reviewed and approved before use.
The study also identified three publicly released inspection records mentioning AI, including Form FDA 483 observations and Establishment Inspection Reports.
None of these additional inspection records cited a firm’s use of AI as a deficiency.
However, the available Form 483 and inspection-report sample was not comprehensive.
What the FDA data tell us
The evidence demonstrates that existing CGMP requirements can be applied directly to activities supported by AI.
It does not demonstrate that AI-related deficiencies are rare across all pharmaceutical manufacturing operations.
Warning Letters represent selected enforcement actions, and the study’s small number of relevant cases does not establish a reliable trend.
United Kingdom: MHRA reports actual AI-generated inspection response failures
The UK’s Medicines and Healthcare products Regulatory Agency (MHRA) has published one of the clearest accounts of real problems associated with AI in regulatory compliance communications.
On 29 June 2026, MHRA inspector Peter Brown described cases involving AI-assisted responses submitted to the agency following GxP inspections.
The problems included:
- References to MHRA guidance that did not exist.
- Incorrect or inappropriate regulatory citations.
- Responses containing significant factual inaccuracies.
- Extensive documentation that failed to address the actual inspection deficiencies.
- Insufficient verification and technical oversight.
One response exceeded 90 pages without adequately addressing the deficiencies.
In another case, an AI-generated response to a serious deficiency affecting patient safety contained significant inaccuracies and non-existent references.
MHRA reported that the time required to review and respond increased from approximately four hours to more than 20 hours.
These are particularly important examples because they document actual failures encountered by the regulator, rather than hypothetical AI risks.
MHRA’s regulatory position
MHRA does not prohibit AI-assisted preparation of inspection responses.
Instead, it expects submissions to be accurate, verifiable, technically reviewed, appropriately approved and supported by evidence.
The agency also offers voluntary disclosure of AI use in inspection responses.
Companies may identify which sections were prepared with AI assistance and confirm human verification and approval.
MHRA warns that inadequate responses may be rejected, returned for correction, or contribute to higher future inspection priority and possible regulatory escalation.
Important scope limitation: The MHRA publication addresses GxP inspection responses generally. It does not provide a separate count of pharmaceutical GMP cases or establish that every example described arose from a manufacturing inspection.
Nevertheless, it demonstrates how unverified generative AI can create tangible regulatory compliance problems.
Switzerland: AI changes can trigger formal notification obligations
Switzerland provides another significant example, but of a different kind.
Rather than reporting an individual AI-related GMP violation, the Swiss medicines inspectorates have explicitly addressed the introduction of AI into pharmaceutical operations.
A technical interpretation document, effective from 21 May 2025, describes major changes requiring notification to Swissmedic under Article 41 of the Medicinal Products Licensing Ordinance.
The examples include introducing or replacing computerised software with AI or machine-learning applications for manufacturing or supply activities.
Notably, the document explicitly mentions:
- LLMs supporting complaint management.
- AI applications for Product Quality Reviews.
- AI-supported batch-record review.
- Machine learning applied to manufacturing processes.
- Automated visual inspection.
- New AI applications in quality control.
This is particularly relevant to GMP because the scope is not restricted to AI controlling manufacturing equipment.
It can also encompass AI applications supporting quality-system activities.
The notification requirement depends on the nature and classification of the change under the applicable Swiss legal framework.
The document does not mean that every experimental AI application automatically requires regulatory notification.
However, it establishes that significant AI-related changes may have external regulatory implications beyond internal validation and change control.
In January 2026, Swissmedic also introduced guidance for Scientific GMDP Meetings, explicitly identifying AI-enabled computer-controlled systems as potential subjects for early regulatory discussion.
Practical implication: For manufacturers operating under Swiss regulatory oversight, AI implementation should include an assessment of external notification requirements as well as internal GMP qualification and validation.
No specific AI-related GMP inspection deficiency was identified in the Swiss publications reviewed for this analysis.
European Union: Annex 22 addresses AI risks before a large enforcement history exists
The European regulatory approach is developing through a dedicated GMP framework.
In July 2025, the European Commission launched consultation on a proposed Annex 22 concerning artificial intelligence in pharmaceutical manufacturing.
The consultation draft focused on static AI models and stated that generative, probabilistic, dynamic and adaptive models should not be used in critical GMP applications.
Following stakeholder feedback, EMA held an expert workshop in June and July 2026.
The official report, published on 2 October 2026, documents discussion of a potentially broader, risk-based approach that could accommodate such technologies under a robust control strategy.
The issues considered included:
- Validation of probabilistic and generative AI.
- Detection and containment of incorrect AI outputs.
- Human oversight and accountability.
- Model monitoring, changes and drift.
- Data integrity, traceability and auditability.
- Supplier oversight and cybersecurity.
Regulatory status: Annex 22 remains a proposal under development. The consultation wording is not an enforceable GMP prohibition, and the workshop report does not establish a final regulatory position.
EU and EEA inspectors also publish statements of GMP non-compliance through EudraGMDP.
However, the public database does not provide a dedicated AI-related deficiency classification.
The available sources reviewed did not establish a confirmed AI-specific GMP non-compliance statement.
This should not be interpreted as evidence that no such observations have occurred.
Individual inspection findings are not generally published in a form that allows comprehensive searching by AI technology or model type.
Japan: explicit treatment of AI models operating under GMP
Japan has made two particularly relevant regulatory publications in November 2026.
PMDA principles - 2 November 2026
The Pharmaceuticals and Medical Devices Agency published its Basic Principles for the Use of AI in Regulatory Activities.
The document covers AI-supported regulatory judgments throughout the pharmaceutical lifecycle, including manufacturing.
Its central themes include:
- Human accountability.
- Risk-based evaluation.
- Clearly defined context of use.
- Data governance and traceability.
- Model evaluation.
- Lifecycle management and monitoring.
The document explicitly states that it does not establish new regulatory requirements.
MHLW CMC model considerations - 5 November 2026
A more directly GMP-relevant document was circulated by Japan’s Ministry of Health, Labour and Welfare (MHLW).
It addresses the use of models, including AI and machine learning, in pharmaceutical manufacturing and quality control.
The document provides a framework for evaluating model credibility according to its intended use, including verification, validation and uncertainty assessment.
One particularly interesting distinction concerns models operating under GMP without directly supporting registered manufacturing or control-strategy information.
Examples include deviation investigation support and internal quality-management tools.
The document indicates that such models may not generally need to be included in the Common Technical Document, but the relevant information should be available for presentation to regulators on request.
It also addresses monitoring of model performance deterioration, updating or retraining models, fallback arrangements and integration with GMP change and deviation management.
The framework is limited to models whose parameters remain fixed during operation. Automatically updating models and models whose behaviour cannot be sufficiently specified or verified are outside its scope.
This exclusion is not equivalent to a general prohibition of such technology.
The document also makes clear that GMP requirements for computerised-system validation and data integrity must be addressed separately.
Practical implication: Japan is developing an increasingly explicit approach to the evidence expected for AI models in pharmaceutical manufacturing and quality-control activities.
However, these publications are regulatory principles and technical considerations, not published inspection observations concerning individual manufacturers.
PMDA also publishes annual GMP inspection reports and anonymised lists of major deficiencies. No confirmed AI-specific GMP case was established from the reviewed public summary material.
Canada: AI-supported compliance work, but limited public inspection detail
Health Canada has described AI initiatives relevant to regulatory inspection and compliance activities.
Its 2026-27 Departmental Plan refers to using image recognition and document analysis to support inspectors assessing regulated products, including drugs.
The department’s 2025-26 results report also describes work assessing these technologies for inspections and compliance verification.
However, these publications do not demonstrate that AI is routinely being used to identify pharmaceutical GMP deficiencies.
Nor do they report a quantified number of GMP violations attributed to pharmaceutical manufacturers’ use of AI.
There is also an important change affecting transparency.
On 26 October 2026, Health Canada announced that public posting of Initial Inspection Deficiencies in its Drug and Health Products Inspections Database had been discontinued effective 21 October 2026.
This applies to drug GMP inspections, among other inspection programmes.
Final compliance outcomes remain publicly reportable.
The change limits the availability of preliminary findings for researchers and complicates attempts to compare historical inspection observations with future data.
[b]Current evidence:
Australia: TGA supports the development of AI-specific GMP guidance
Australia’s Therapeutic Goods Administration (TGA) has participated in international work on modernising GMP requirements.
On 26 August 2025, TGA published a communication concerning proposed revisions to the PIC/S GMP Guide.
The proposed changes include a new Annex 22 dedicated to artificial intelligence.
TGA highlighted the importance of updating GMP guidance as AI and other digital technologies enter pharmaceutical manufacturing.
Australia applies the PIC/S GMP Guide within its regulatory framework, making future adoption of AI-related revisions particularly relevant to local manufacturers.
However, the proposed Annex 22 is not automatically an effective Australian GMP requirement merely because it has been circulated for consultation.
The reviewed public TGA publications did not identify a specific GMP deficiency attributed to improper AI use in medicinal-product manufacturing.
General observations about deficient computerised systems, validation or data integrity should not automatically be classified as AI-related.
China: NMPA plans AI-supported inspections and manufacturing-risk surveillance
China’s National Medical Products Administration (NMPA) has set out a broad strategy for AI in pharmaceutical regulation.
Its Implementation Opinions on Artificial Intelligence + Drug Regulation, issued in 2026 and made available in English in August, describe plans to expand AI capabilities across regulatory activities.
The strategy includes:
- AI-supported regulatory risk assessment.
- More intelligent inspection planning.
- Automated identification of potential compliance signals.
- Analysis of manufacturing-related information.
- AI-supported monitoring of high-risk products.
- Greater use of regulatory data and intelligent agents.
The document also addresses model reliability, validation, transparency and information security.
Its ambition extends to 2030 and beyond.
However, this is a regulatory development and implementation strategy.
It does not demonstrate how many GMP inspections currently use AI, nor does it report the number of manufacturers cited for inappropriate AI applications.
No individually documented AI-related pharmaceutical GMP deficiency was confirmed in the reviewed public NMPA material.
Brazil: AI results may contribute to GMP certification risk assessment
Brazil’s Anvisa introduced an explicit reference to artificial intelligence in RDC 982/2025, dated 28 July 2025.
The resolution addresses risk-management criteria relevant to issuing or renewing Good Manufacturing Practices Certificates.
Article 3, paragraph 3, allows results produced by AI models to be used as risk-management criteria when applicable technical requirements are met.
These include technical consistency, data protection, traceability and validation.
The resolution also establishes conditions for applying risk-management mechanisms in certification decisions.
Importantly, its implementation is subject to the approvals and procedures specified in the regulation.
[b]The distinction is significant:
This does not prove that AI is already used in every GMP certification decision.
Nor does it establish an AI-related GMP violation by a pharmaceutical manufacturer.
The reviewed Anvisa materials did not provide a publicly verified count of such violations.
Other authorities: Singapore, South Korea and Poland
Singapore - Health Sciences Authority (HSA)
HSA publishes annual summaries of GMP inspection deficiencies, including reports for 2023, 2024 and 2025.
These reports provide information about conventional GMP deficiency categories, such as pharmaceutical quality systems, premises, production, documentation and quality control.
However, AI is not presented as a separate, consistently reportable GMP deficiency category.
The reviewed public reporting therefore does not support a reliable count of AI-specific findings.
South Korea - Ministry of Food and Drug Safety (MFDS)
MFDS recognises AI and digital technologies as important regulatory developments.
Its public materials include initiatives concerning AI-related regulatory modernisation and guidance for AI-enabled medical devices.
However, medical-device AI guidance must not be mistaken for pharmaceutical GMP inspection findings.
No specific publicly documented pharmaceutical GMP deficiency caused by the use of AI was confirmed in the reviewed MFDS sources.
Poland - Chief Pharmaceutical Inspectorate (GIF)
GIF conducts inspections of pharmaceutical manufacturers and participates in the European GMP regulatory framework.
It also carries out inspections outside Poland and contributes to the European system of GMP certification and compliance oversight.
In the public GIF sources reviewed, no specific inspection observation explicitly addressing a pharmaceutical manufacturer’s use of AI was confirmed.
This does not establish that Polish inspectors have never encountered AI-related problems.
Inspection reports are not comprehensively published as searchable, full-text records, and relevant findings may be recorded under existing GMP categories.
Can we measure a global trend in AI-related GMP deficiencies?
Not reliably from the currently reviewed public sources.
The available evidence is highly uneven.
The FDA study provides a defined document population, an explicit search methodology and reproducible counts.
MHRA provides real examples of AI-related compliance failures but no total number of affected inspections or submitted responses.
Other regulatory authorities publish inspection statistics, non-compliance statements, policy documents or general GMP deficiency categories without consistently identifying AI as a distinct subject.
There are also important differences in the documents being compared.
- A Warning Letter is not the same as a routine inspection observation.
- An incorrect response to a GxP inspection is not necessarily a manufacturing GMP deficiency.
- A GMP non-compliance statement may summarise several underlying deficiencies.
- A draft AI guideline is not evidence that a manufacturer violated a requirement.
- A regulator’s internal use of AI is not evidence of industry non-compliance.
- A missing publicly accessible case is not proof that no case exists.
Even basic terminology creates challenges.
Regulators may describe AI-related applications as machine learning, predictive analytics, automated algorithms or computerised systems.
Conversely, terms such as “AI” may have unrelated meanings in pharmaceutical documentation, including “active ingredient” in certain regulatory contexts.
A meaningful international study would therefore require document-level review, careful classification and consistent inclusion criteria.
The central finding is not that the world has recorded only one AI-related GMP violation. It is that the public evidence is currently too fragmented to calculate a defensible global total.
What pharmaceutical companies can learn from this comparison
Although the available evidence does not establish a global enforcement trend, several practical regulatory lessons are emerging.
1. Existing GMP responsibilities remain applicable
The FDA case demonstrates that AI-generated documentation remains subject to established Quality Unit responsibilities.
2. AI-generated regulatory communications must be verified
The MHRA examples show that plausible but incorrect AI-generated information can create immediate compliance consequences.
3. AI implementation may require external regulatory assessment
The Swiss framework demonstrates why manufacturers should evaluate notification obligations when introducing significant AI-enabled systems or modifying established processes.
4. Validation should reflect the intended use and model risk
The Japanese and European publications emphasise context-specific verification, validation, uncertainty, human oversight and lifecycle controls.
5. Regulators are developing their own AI capabilities
Developments in China, Brazil, Canada and the United States indicate increasing interest in AI-supported regulatory analysis and inspection prioritisation.
Manufacturers should not assume that these initiatives already provide proven, comprehensive automated GMP surveillance.
Nevertheless, they demonstrate a direction of regulatory development that should be monitored.
AIforPharma assessment: regulation is moving faster than published enforcement evidence
The comparison reveals an interesting asymmetry.
Pharmaceutical regulators are developing increasingly detailed expectations for AI used in manufacturing, quality control and regulatory documentation.
At the same time, publicly available enforcement material contains relatively few cases explicitly identifying AI as the source of a GMP failure.
Several explanations are possible.
AI adoption in critical GMP applications may still be limited. Regulators may address AI-related deficiencies using established GMP requirements without specifically mentioning the underlying technology. Inspection reports may also remain confidential or be published only in abbreviated form.
The available evidence cannot determine the relative importance of these explanations.
However, three examples already reveal different dimensions of future AI compliance oversight:
- FDA: Are AI-generated GMP outputs appropriately reviewed and approved?
- MHRA: Are AI-assisted regulatory submissions accurate, relevant and supported by genuine evidence?
- Swissmedic: Has the company appropriately assessed and notified significant changes involving AI-enabled systems?
Japan and the EU add further questions concerning model credibility, lifecycle management and risk-based controls.
These are not necessarily new standalone GMP obligations.
Rather, they illustrate how existing responsibilities and emerging regulatory expectations may apply to increasingly sophisticated technologies.
[b]The most important conclusion is that pharmaceutical companies should not wait for a large number of AI-related Warning Letters or inspection findings before establishing appropriate governance.
The absence of extensive published enforcement evidence should not be interpreted as regulatory acceptance of uncontrolled AI use.
For pharmaceutical quality professionals, the priority remains straightforward: ensure that AI applications are suitable for their intended purpose, supported by appropriate evidence, governed by effective controls and subject to accountable human oversight.
Scope, methodology and limitations
This article is a targeted comparative documentary analysis of publicly accessible regulatory sources available as of 7 November 2026.
It covers the FDA, MHRA, European medicines regulatory network, Swissmedic, PMDA/MHLW, Health Canada, TGA, NMPA, Anvisa, HSA, MFDS and GIF.
The analysis distinguishes:
- Published enforcement actions and identifiable inspection-related problems.
- Regulatory guidance, technical interpretations and official positions.
- Draft guidance and consultation documents.
- Regulatory strategies and internal AI adoption.
- Situations in which AI-specific GMP observations could not be established from the reviewed public sources.
The FDA quantitative results are drawn from a published retrospective scientific study, not from an independent repeat analysis of all 3,011 Warning Letters.
For other jurisdictions, the review was a targeted assessment of authoritative public sources, not an exhaustive full-text search of every inspection report.
Consequently, the analysis does not provide comparable national incidence rates, a worldwide deficiency total or a statistically validated trend.
Statements that no AI-specific case was identified refer only to the public material reviewed.
Official sources and publication dates
United States - FDA
- Kwiecinski G., Nandy R. Artificial Intelligence in FDA Regulatory Correspondence, 2022 to 2026: A Retrospective Document Analysis. Discover Pharmaceutical Sciences. Peer-reviewed scientific paper, published 31 October 2026. Open access:
https://doi.org/10.1 … 7/s44395-026-00061-5 - FDA - Purolea Cosmetics Lab, Warning Letter 320-26-58, issued 2 April 2026:
https://www.fda.gov/ … -lab-722591-04022026
United Kingdom - MHRA
- MHRA Inspectorate - Use of AI for GXP Inspection Responses: Setting Standards Without Stifling Innovation. Inspectorate publication, 29 June 2026:
https://mhrainspecto … stifling-innovation/
Switzerland - Swissmedic
- Swiss medicines inspectorates - Notifications According to Article 41 MPLO. Technical Interpretation I-SMI.TI.22e, version 3.0, effective 21 May 2025:
https://www.swissmed … ions_Art_41_MPLO.pdf - Swissmedic - Scientific GMDP Meetings for Establishment Licence Holders. Guidance document, effective 6 January 2026:
https://www.swissmed … _BBW-Inhaber_V01.pdf
European Union - EMA / European Commission
- European Commission - Consultation on EU GMP Chapter 4, Annex 11 and New Annex 22. Consultation opened July 2025 and closed 7 October 2025:
https://health.ec.eu … s-chapter-4-annex_en - EMA - Multistakeholder Workshop on AI Guidance Development (Annex 22). Official workshop report, published 2 October 2026:
https://www.ema.euro … ment-annex-22_en.pdf - EMA - EudraGMDP public GMP non-compliance database:
https://eudragmdp.em … hGMPNonCompliance.do
Japan - PMDA / MHLW
- PMDA - Basic Principles for the Use of AI in Regulatory Activities (Early Consideration). Official principles, 2 November 2026:
https://www.pmda.go. … /files/000283394.pdf - MHLW - Considerations for the Use of Models, Including AI and Machine Learning, in Pharmaceutical Manufacturing and Quality Control (CMC). Official communication with technical considerations, 5 November 2026; Japanese and English text:
https://www.mhlw.go. … chi/T261105I0010.pdf - PMDA - Quality Assurance Activities and GMP/GCTP Annual Reports, including FY 2025 report issued in English on 2 October 2026:
https://www.pmda.go. … p-qms-gctp/0007.html
Canada - Health Canada
- Health Canada - 2026-27 Departmental Plan, published 2026:
https://www.canada.c … partmental-plan.html - Health Canada - 2025-26 Departmental Results Report, available November 2026:
https://www.canada.c … -results-report.html - Health Canada - Discontinuing the Public Posting of Initial Inspection Deficiencies. Drug Establishment Licensing Bulletin 197, published 26 October 2026:
https://www.canada.c … ctions-database.html
Australia - TGA
- TGA - Have Your Say on Updates to GMP Guidelines. Official consultation announcement, published 26 August 2025:
https://www.tga.gov. … ctice-gmp-guidelines
China - NMPA
- NMPA - Implementation Opinions on Artificial Intelligence + Drug Regulation. Official policy document issued in 2026; English publication 20 August 2026:
https://english.nmpa … -08/20/c_1210674.htm
Brazil - Anvisa
- Anvisa - RDC 982/2025. Official resolution dated 28 July 2025; English translation:
https://www.gov.br/a … s/rdc-982-2025-e.pdf
Singapore - HSA
- HSA - Good Manufacturing Practice and Good Distribution Practice Standards. Official resource containing annual GMP deficiency trend reports for 2023, 2024 and 2025:
https://www.hsa.gov. … r-wholesale/gmp-gdp/
South Korea - MFDS
- MFDS - 2026 Agency Work Plan, dated 16 December 2025 (Korean-language official publication), covering regulatory modernisation and emerging AI technologies:
https://www.mfds.go. … /etc/2026_mfds_1.pdf