Innovating Regulation: Global Perspectives on Modernizing Quality and Good Manufacturing Practices (GMP)
A panel discussion at the 2026 ISPE Europe Annual Conference brought together leading regulators from Europe and the US to explore how regulatory leadership is evolving in an era of rapid technological and scientific change, including digital transformation efforts, GMP modernization, the maturation of quality risk management, and introduction of artificial intelligence (AI). The panel also addressed the challenges of global convergence, and the need for dynamic collaboration between industry and regulators.
The panelists were:
The session was moderated by Alice Redmond, Chief Risk and Quality Officer, CAI.
Question 1: In the context of accelerating scientific and technological innovation, how can regulatory frameworks remain sufficiently agile, and what concrete role should industry play in shaping and sustaining that adaptability?
Korakianiti started by emphasizing that innovation is a major strategic initiative for the European Union (EU). Early engagement with regulators is very important and she reminded the audience that EMA’s door is open. She recognized that there is work to prepare, and you may not receive the answers you would like; however, this is a process for the industry to de-risk a project. Industry should focus on telling the story of how a technology impacts the product and patient, not just discussing the technology. Industry should also publish their technology—regulators read publications! Finally, regulators do establish pilot programs, and industry should have no hesitation in volunteering for them. Use of pilots is a mechanism for both regulators and industry to learn and understand one another.
Hoefnagel emphasized that expertise lies with industry, and the industry needs to educate regulators. If you can demonstrate that you understand the product, the process, and the risks and explain them thoroughly, you can do anything.
Cuddy described the large effort of regulators to set up pilots, for example, the EMA pharmaceutical quality system (PQS) effectiveness pilot, and stressed the EMA’s need to learn. His message was “please engage and be ready to accept failure.” Korakianiti added that we are all experimenting, and many experiments end in failure. She also commented that we do need to learn.
Jackson supported the theme of pilots and referred to the United Kingdom’s strategy to create a pro-innovation regulatory environment, for example, using the MHRA Innovation Office.
Fisher emphasized the US FDA’s commitment to supporting innovation, which commenced more than a decade ago with the Emerging Technology Program (ETP). This initiative by the US FDA was the first opportunity for technology developers to approach the agency, early so that technical and regulatory issues could be identified and hopefully resolved before filing. Most people assumed that the discussions would focus on scientific and technical matters; however, these accounted for about half of the discussions, the other half being on regulatory policy issues. The US FDA then decided to create a sister initiative, Framework for Regulatory Advanced Manufacturing Evaluation (FRAME), that can address some of these issues and take a hard look at the regulatory framework and how it relates to advanced technologies. Hence the FRAME initiative. Fisher agreed with other panelists that the industry, including suppliers, is the expert. A key goal of FRAME is harmonization, hence US FDA consults EMA on many guidances during drafting, and has participated in several EMA Quality Innovation Group (QIG) meetings, including Listen and Learn sessions in the presence of industry presenters. This interaction has been found extremely helpful. One of the four technology focus areas for the FDA is distributed manufacturing, for which there has been stakeholder feedback and two draft guidances on GMP considerations.
Question 2: The issuance of the 10 common principles on the use of AI in pharmaceutical industry by EMA and US FDA in January 2026 is a notable collaboration. Will this be the starting point for further regulation?
The Draft Annex 22 on artificial intelligence (AI) has been a major subject for discussion at the conference, and Cuddy pointed out that it broadly aligns with the 10 guiding principles of good AI practice in drug development issued jointly by EMA and US FDA. Annex 22 was drafted in cooperation with the Pharmaceutical Inspection Cooperation Scheme (PIC/S), and two US FDA colleagues serve on the drafting group, which emphasizes the strength of collaboration.
Korakianiti indicated that, for the EU as for the US FDA, strong collaboration is very important. This collaboration is demonstrated at QIG meetings and Listen and Learn sessions. There is also information sharing on new and forthcoming submissions with the US FDA.
Fisher was strongly supportive of collaboration and mentioned the 2025 ICH Reflection Paper, Proposed ICH Guideline Work to Facilitate the Adoption of Advanced Pharmaceutical Manufacturing. This paper identifies key areas where globally harmonized regulatory guidelines could facilitate the adoption and regulation of advanced manufacturing technologies to improve pharmaceutical development and manufacturing. Examples of advanced manufacturing technologies, such as process modelling, including AI-based models, and decentralized or distributed manufacturing, would benefit from further ICH activity. ICH guidance on process modelling could extend beyond that discussed in the ICH Q8/Q9/Q10 Implementation, Points to Consider document, 2011. Fisher stressed that ICH involves industry as well as regulators.
Question 3: What limits the use of first‑tier regulatory decisions and harmonized data to avoid repeat approvals across regions?
Jackson referred to the importance of the International Coalition of Medicines Regulatory Authorities (ICMRA) pilot programs focusing on i) collaborative assessments of chemistry, manufacturing, and control (CMC) related post-approval changes and ii) hybrid inspections. Learnings were reviewed at a workshop in August 2025. The pilot is testing regulatory reliance and convergence. Jackson stressed that there are lots of learnings, including issues identified by the industry so regulators can evaluate how to move forward. Many of the issues are challenging, and some may require a change in legislation.
Cuddy referred to the EMA Inspectors Working Group (IWG)'s successful pilot, which evaluated IWG members accepting PIC/S members’ inspection reports. He also mentioned that the new EU Pharmaceutical Legislation does allow EU members to accept others’ inspection reports.
Korakianiti did recognize that ICH Guidelines could have some differing interpretations during implementation. These issues need to be worked on; however, there are priorities with the EU focusing on innovative technology and resource consideration, since a different way of working does require more effort. Key identified issues to move forward are a common IT platform and data structuring. Nonetheless, the industry could participate in ICMRA pilots now and gain benefit.
Question 4: As digital ecosystems expand to include cloud platforms, internet of things (IoT) sensors, and real-time release testing, how do regulators envision evolving expectations for data integrity, validation, and cybersecurity without constraining technological innovation?
Hoefnagel referred to work on harmonized data standards, the use of open architecture software, and the introduction of common IT platforms. The work of ICH and the International Organization for Standardization (ISO) on these issues is to be commended.
O'Donnell indicated that current GMPs support use of innovative technologies since they have included risk-based considerations for many years, e.g., EU GMP Annex 11 (Computerized Systems) and Chapter 4 (Documentation), and the new AI draft Annex 22. These risk-based GMPs are less prescriptive and enable flexible interpretation and support the introduction of new technology. Risk-based approaches inform data integrity and security as well as cybersecurity. O'Donnell complimented ISPE on their risk-based guide on commissioning and qualification, and mentioned that the forthcoming ISPE Guide on quality risk management was aligned with ICH Q9, (R1) revision. ICH Q9(R1) revision and training materials mention that the application of digitalization and new technology can lead to risk reduction.
Question 5: How do regulators foresee AI changing the nature of engagement with industry across scientific advice, submissions, inspections, and lifecycle oversight, and what new competencies or transparency will be expected from companies?
Fisher reflected that AI is broadening with whom CDER/CBER. For example, the recent US FDA Guidance on AI, Considerations for the Use of Artificial Intelligence to Support Regulatory Decision-Making for Drug and Biological Products, includes both clinical and CMC considerations. US FDA has heard that some companies are reluctant to discuss AI applications, however, Fisher stressed that US FDA has a proven track record of engagement on new technologies and companies should not hesitate to come forward. A recent FDA warning letter cited inappropriate use of AI. The main message from this is that there is need for human intervention, overview, and ultimate responsibility. Cuddy pointed out that normal GMP still applies.
Korakianiti mentioned that site visits are very helpful to make proposals real, and companies should suggest these, accepting that it may be more work; however, there are benefits to de-risk a project.
Question 6: Pharmaceutical quality system (PQS) effectiveness is currently very topical in the European Economic Area (EEA), with the 12-month inspections pilot launched by EMA in March 2026. Why did this come up now, given that PQS effectiveness has been a GMP requirement for many years?
O'Donnell indicated that there is a lot for companies to examine in terms of PQS effectiveness, particularly the fitness of purpose of their change management system and lifecycle management generally. For example, how is quality risk management (QRM) applied, and what impact does use of AI in GMP applications have? It is highly likely that AI models will require updating with use and time, and also that current change management systems will also require updating. AI updating is not a simple risk assessment exercise. O'Donnell indicated that Q9(R1) describes the application of the concept of formality, which requires considerations of the level of uncertainty, importance, and complexity. These considerations are very applicable to understanding the level of risk to an AI change management situation. For example, different contexts of use of AI will have different levels of importance for influencing and informing decision-making, and different AI models will have different levels of complexity. The level of complexity of AI applications may differ between applications.
He continued that the use of AI is likely to impact how companies conduct management review. Using AI, data and information will need to be presented in new ways for management review meetings, whereby senior management can look at the risks presented, and what risk mitigation actions are proposed.
As a reviewer, Hoefnagel said that in a science- and risk-based submission, it is necessary to understand PQS effectiveness. Regarding the use of AI, it is essential that reviewers comprehend the performance of AI, what it does, and what its objective is, and how the lifecycle, for example, the drift of an AI application, is managed. If AI is used to create regulatory documents, these need to be read very carefully to ensure they are accurate and make sense.
Cuddy reminded the audience of the EMA PQS effectiveness pilot mentioned earlier and requested companies to engage.
Question 7: There has been a lot spoken about quality culture in recent years, but there is not much about it in official regulatory guidance. How important do you think quality culture is in an environment of rapidly advancing scientific and technical innovation?
O'Donnell commented that the topic has many angles. An organization’s culture underpins the benefits and risks to introducing new technology. It impacts the level of risk tolerance. Poor behavior would be evident when a risk assessment exercise was performed to meet pre-determined outcomes, which undermines the purpose of conducting a QRM exercise. Staff recruitment, training and management commitment and communication drive good behaviors. Leadership is very important.
Cuddy reflected that issues with data integrity in the past have been due to poor culture. Failure has been talked about, and how to deal with failure is impacted by culture. Does an organization fire an individual or does it learn from a failure? It is top management responsibility. He asked the audience if quality culture should be mentioned in regulations. The ISPE audience and regulators generally agreed that formal regulation of quality culture would be counterproductive. Quality culture consists of values, behaviours, and mindsets, driven by company leadership throughout the company and regulators already address culture indirectly through inspection outcomes.
Korakianiti referenced the ten principles of behavior expected of employees that were on the wall of the Carlsberg brewery placed there 150 years ago and are applicable to a company today.
Fisher highlighted the US FDA quality maturity management program, which has been running for a number of years. There have been two pilots with about 20 establishments involved. US FDA has been promoting this program as one that leads to benefits for an organization as described in a white paper, Quality Management Initiatives in the Pharmaceutical Industry: An Economic Perspective published in July 2025. He commented that there was evident satisfaction with some companies requesting additional sites join the program. This program is continuing to a third year.
Jackson commented that culture will be important in decentralized manufacturing, where the role of the qualified person releasing batches from the network will be to ensure that constituent parts of the network meet requirements.
Finally, Redmond pointed out the extensive work has been performed by ISPE on quality management maturity and culture, for example, the Advancing Pharmaceutical Quality program, which includes a Guide on Cultural Excellence.
Question 8: What stage is the Q9 revision work at now, and can you tell us a little about the newly issued QRM training materials from ICH? ICHQ9R1 is in place since January 2023 with another recent release of training materials; what impact are you seeing from an inspection perspective and what does the industry need to focus on to drive implementation and maturity?
As regulatory leader and ICH Q9R1 Rapporteur of the ICH Q9(R1) revision, O'Donnell replied that he has seen different actions by companies in response to the publication of the revision and the training material in 2023. Some companies are taking action and using the ICH training materials, whilst many others seem to be doing little. Further training material was added to the ICH website in 2024, and the final tranche in March 2026. All this material is now easier to find on the ICH website. He pointed out that the concepts expanded in the revision are difficult, these being subjectivity, formality, risk-based decision-making, and hazard identification. Given these challenges, he recommended that the industry share its learnings, for example, best practices of implementing the difficult concepts or using a particular QRM tool.
Question 9: With the revised EU Annex 1 now embedded, what are regulators’ expectations for demonstrating ongoing maturity and continual improvement in contamination control strategies (CCS), QRM implementation for aseptic manufacturing beyond initial compliance?
Jackson stressed, and Cuddy agreed, that the CCS should be a living document, subject to continual improvement as issues are identified from, for example, deviations, out of specifications results, changes and updated risk assessments. It is necessary for organizations to learn and change, updating the CCS, and this learning is at the heart of management review. CCS improvement is linked to PQS continual improvement and PQS effectiveness. Jackson referred to a draft PIC/s Q&A document on inspector learnings from Annex 1 implementation, which is being finalized.
Question 10: Beyond environmental management systems, how do regulators view the role of product specific sustainability considerations, such as solvent reduction or substances of concern, in regulatory submissions and lifecycle decision making?
Hoefnagel mentioned that this topic was the subject of a recent EMA QIG Listen and Learn Focus Group meeting. There are many sustainability topics that cross between products and companies, such as solvent reduction and the replacement of PVC blisters. He recommended that companies should collaborate and share learning and data. Regulators and industry need to find more efficient ways of working since many of the sustainability topics require post-approval regulatory filing and are needed quickly. Ways must be found to facilitate multi-product filings.
Korakianiti recognized the importance of sustainability with many topics linked to cost reduction or reduction in animal testing. She agreed that there will be a huge burden on industry and regulators with the large number of post-approval applications, and consequently, new regulatory processes may be required. Currently, the industry should engage early with regulators regarding bundling company initiatives and agreeing work packages as well as regulatory pathways, for example, using reliance via the World Health Organization (WHO) program. This engagement process is difficult; however, working together, it should be possible to use existing tools to facilitate global introduction.
Question 11: What is one action regulators or the industry should take in the next 12-24 months?
Cuddy would like to see the EMA IWG deliver the 3 year work plan, which includes EU GMP Chapters 1 (pharmaceutical quality system) and 4, and Annexes 11 and 22. There is a large amount of work given the number of comments on draft Annex 11, and the challenge of Annex 22 with the upcoming industry feedback workshop. Additionally, updating the EudaGMP database would benefit industry and regulators globally.
Korakianiti indicated that operationalizing collaborative regulatory assessments should facilitate faster global introduction of new technology, which has a compelling justification. She would also like to see the progression of a global data standard and ICH M16, Structured Quality Product Submissions, which is an agreed ICH topic. The extension of the use of the EMA centralized procedure should lead to faster introduction of new products in the EU.
Hoefnagel agreed on the need to progress the use of structured data in, for example, submissions. Data standards may allow submissions with reduced data, with the use of digital tools to assess products post approval. Implementation of the new EU pharma legislation should facilitate more use of decentralized manufacturing, not just for advanced therapy medicinal products (ATMPs). We need to understand what “platforms” are and how they could be used to facilitate faster submission and approval.
Jackson stressed he would like more of what we have been discussing during the session—more sharing, collaboration, engagement, and learning. We should embrace AI with care. Recognition is also important for those delivering and supporting major actions.
Fisher would like to see the ICH reflection paper, Proposed ICH Guideline Work to Facilitate the Adoption of Advanced Pharmaceutical Manufacturing, October 2025, progressed, which maps out all ICH would like to strategically deliver in the next period. Fisher particularly would like the topic of process modelling advanced, which is an umbrella for AI applications applied to process understanding.
O'Donnell emphasized that he would like to work together to understand the risks of using AI models. Better application of QRM and learning generally, and how best to use knowledge, would benefit all parties.
Concluding Remarks
Redmond summarized the messages from the session as communication, collaboration, working together, starting early, not being afraid to fail, getting involved in the pilots, and just going for it. She thought the industry has, or had, an inherent level of cautiousness in approaching regulators. She considered that was the past, and she thought it was changing due to forums such as this. Industry should engage and become involved in pilots. The effort to comment on documents on which regulators have spent a lot of time collaborating is worthwhile, so that the voice of the industry and the voice of the regulators come together.
Redmond expressed a huge thank you to the panelists for their fantastic contributions to this session, as well as to the audience for their attention.
Learn more about the ISPE Guide: Quality Risk Management
Disclaimer
This is a summary of a presentation made on 22 April at the 2026 ISPE Europe Annual Conference in Copenhagen, Denmark. It has not been vetted by any of the regulators or agencies mentioned in this article, nor should it be considered the official positions of the agencies mentioned.
