Pharmaceutical manufacturing increasingly relies on computerized systems, electronic records, automated processes, and digital technologies. These systems can improve efficiency and traceability, but they also introduce risks that must be controlled to protect product quality, patient safety, and data integrity.
The European Commission has consulted on revisions to EU GMP Chapter 4 (Documentation), Annex 11 (Computerised Systems), and a new Annex 22 (Artificial Intelligence). The proposed framework focuses on strengthening documentation controls, computerized-system lifecycle management, data governance, and the use of AI in pharmaceutical manufacturing. The consultation closed on 7 October 2025. Organizations should check the latest official publications before treating proposed provisions as effective requirements.
For pharmaceutical manufacturers and GxP service providers, this is an opportunity to review existing controls, identify potential gaps, and prepare for evolving regulatory expectations.
1. Chapter 4: Documentation and Data Governance
Chapter 4 addresses documentation within the pharmaceutical quality system. Documentation must provide reliable evidence that activities have been performed in accordance with approved procedures and applicable GMP requirements.
The proposed revision places greater emphasis on risk-based data governance and the management of documentation across paper, electronic, and hybrid environments.
During a review, organizations should consider:
- Whether records are complete, accurate, legible, and retrievable
- Whether documents are controlled, approved, and maintained throughout their lifecycle
- How electronic records and signatures are managed
- Whether changes and corrections are traceable
- Whether records remain available and readable for the required retention period
- Whether responsibilities for document management are clearly defined
What to do now: Review document-control procedures, electronic-record practices, record-retention arrangements, and controls for hybrid documentation systems. Identify gaps and prioritize improvements based on risk.
2. Annex 11: Computerised Systems
Annex 11 addresses computerized systems used in GMP-regulated activities. These systems may include Laboratory Information Management Systems (LIMS), Manufacturing Execution Systems (MES), Chromatography Data Systems (CDS), electronic batch records, and other systems that support GxP processes.
The proposed revision emphasizes lifecycle management, quality risk management, supplier oversight, data integrity, audit trails, electronic signatures, and system security.
Key areas to review include:
- System validation: Evidence that the system is fit for its intended use
- Risk assessment: Identification and control of risks to product quality and data integrity
- Access management: Appropriate user permissions and accountability
- Audit trails: Protection, availability, and appropriate review of relevant system activity
- Supplier oversight: Assessment and management of external service providers
- Change control: Evaluation and approval of system changes
- Periodic review: Confirmation that the system remains suitable and controlled
- Data management: Reliable storage, retrieval, retention, and protection of GxP records
What to do now: Prepare an inventory of GxP systems, review validation documentation, assess user-access controls and audit-trail practices, and verify that system changes and suppliers are managed through appropriate quality procedures.
3. Annex 22: Artificial Intelligence in GMP
The proposed Annex 22 focuses on the use of artificial intelligence (AI) and machine learning in pharmaceutical manufacturing. Its proposed approach addresses the selection, training, validation, intended use, performance monitoring, and ongoing oversight of AI models.
AI applications may support activities such as process monitoring, visual inspection, trend analysis, and decision support. However, their use in GxP environments requires careful assessment of how model outputs may affect product quality, manufacturing decisions, or compliance.
Organizations considering AI in GxP activities should evaluate:
- Intended use: What the AI system is designed and permitted to do
- Data quality: Whether training, testing, and operational data are appropriate and reliable
- Model performance: How accuracy and limitations are evaluated
- Validation and documentation: Evidence that the system performs as intended
- Change management: How model updates and changes are assessed and controlled
- Ongoing monitoring: How performance drift or unexpected behaviour is detected
- Human oversight: When review or intervention by qualified personnel is necessary
- Risk management: Potential consequences of incorrect or unreliable outputs
What to do now: Identify existing or planned AI applications used in GxP activities, assess their potential impact, document their intended use, and establish suitable governance and oversight before relying on their outputs for quality-critical decisions.
4. A Practical Readiness Plan
Organizations do not need to wait for every regulatory detail to be finalized before reviewing their current systems. A structured readiness assessment can help identify improvement opportunities without assuming that draft provisions are already binding.
| Step | Recommended action |
|---|---|
| 1. Identify | List relevant documentation processes, computerized systems, and AI applications used in GxP activities. |
| 2. Assess | Compare current controls with applicable requirements and the proposed revisions relevant to each activity. |
| 3. Prioritize | Evaluate gaps according to potential impact on product quality, patient safety, and data integrity. |
| 4. Plan | Assign responsibilities, timelines, and resources for agreed improvements. |
| 5. Document | Record assessments, decisions, actions, and supporting evidence through the quality system. |
| 6. Monitor | Review official regulatory publications and update the readiness plan when necessary. |
5. Common Areas of Concern
A readiness review may identify issues such as:
- Incomplete validation documentation
- Weak user-access or audit-trail controls
- Unclear responsibilities for computerized-system suppliers
- Inadequate management of electronic records
- Insufficient oversight of system changes
- AI applications operating without a clearly documented intended use
- Limited monitoring of system or model performance
- Procedures that have not kept pace with changes in technology
These are potential areas for assessment, not automatic evidence of non-compliance. Each issue should be evaluated against the applicable requirements and the organization's actual processes.
Conclusion
The proposed revisions to Chapter 4, Annex 11, and Annex 22 highlight the importance of reliable documentation, controlled computerized systems, strong data governance, and appropriate oversight of AI in pharmaceutical manufacturing.
Organizations can prepare by understanding their systems, reviewing existing controls, assessing risks, and documenting practical improvement plans. However, the applicable regulatory position should always be confirmed against the latest official European Commission publications.
PharmaTrak's independent GxP audit services can help organizations examine documentation practices, computerized-system controls, data-integrity arrangements, and related quality-system processes to identify gaps and support informed compliance decisions.
Regulatory note: This article is for general educational purposes. It discusses proposed regulatory revisions and readiness considerations; it should not be interpreted as confirmation that draft provisions are in force. Always verify the current status and applicable text of EU GMP requirements before making compliance decisions.



