Upcoming 2026 ISPE Annual Meeting & Expo Session, Autonomous Quality Systems – AI-Enabled Compliance, Covered by Pharmaceutical Online
28 September, 2026
As pharmaceutical companies face mounting pressure to validate increasingly complex digital systems with limited resources, artificial intelligence is emerging as a valuable tool for streamlining validation activities. The topic will be a focal point at the 2026 ISPE Annual Meeting & Expo, where Louie Rayal, Vice President of Governance, Risk, and Compliance at GSK, will discuss how AI can support validation programs while preserving the risk-based decision-making and human accountability required in regulated environments.
According to Rayal, validation teams today must manage expanding digital ecosystems that include cloud-based platforms, AI-enabled applications, automated systems, and increasingly frequent software updates. At the same time, regulatory expectations surrounding product quality, patient safety, and data integrity remain as stringent as ever. This combination has created pressure to deliver validation activities more efficiently without compromising compliance. Rayal believes that the US Food and Drug Administration’s Computer Software Assurance (CSA) framework, introduced in 2022, combined with modern AI capabilities, can help organizations focus resources on activities that have the greatest impact on quality and patient safety.
One of the most significant benefits of AI in validation is its ability to reduce the burden of documentation. Traditional validation programs often generate lengthy narrative documents, extensive test scripts, and detailed traceability records that can consume substantial time without necessarily improving quality outcomes. Rayal argues that AI can help generate concise summaries, identify gaps, and link evidence across requirements, risks, and testing. By aligning with CSA principles, organizations can focus less on producing excessive paperwork and more on sharing objective evidence that systems perform as intended. Importantly, essential artifacts such as risk assessments, test results, approvals, and traceability records remain intact, while redundant documentation is minimized.
The article also explores AI’s growing role in risk management activities, particularly failure mode and effects analysis (FMEA). Traditionally, FMEA exercises require extensive cross-functional workshops and significant institutional knowledge to identify potential risks. Rayal believes AI can accelerate this process by analyzing historical information, industry practices, and prior implementations to suggest potential failure modes that validation teams might otherwise overlook. However, he emphasizes that AI should augment rather than replace expert judgment. Subject matter experts remain responsible for determining which risks are credible, evaluating their potential impact, and deciding what controls should be implemented. AI broadens the conversation, but humans remain the final decision-makers.
Maintaining that balance between automation and human judgment is especially important given the expectations outlined in ICH Q9 (R1), which promotes critical thinking and risk-based decision-making rather than checklist-driven compliance. Rayal acknowledges concerns that some organizations may become overly reliant on AI-generated recommendations. To address this risk, he advocates for strong governance frameworks that position AI-generated outputs as starting points requiring expert review rather than authoritative conclusions. Under this model, accountability for validation and quality decisions remains with qualified personnel, ensuring that AI enhances rather than diminishes critical thinking.
Throughout the validation lifecycle, AI offers value in a variety of administrative and analytical tasks. Rayal identifies draft risk assessments, test scenario development, traceability reviews, validation report preparation, and results summarization as areas where AI can significantly reduce manual effort. In contrast, decisions requiring business context, technical expertise, and regulatory accountability continue to depend heavily on human involvement. Activities such as defining intended system use, determining system criticality, assessing risk acceptance, investigating deviations, and approving final release decisions remain firmly within the domain of qualified professionals.
Modern testing environments illustrate this partnership between humans and AI particularly well. Many software testing platforms now incorporate AI capabilities that can automatically generate and execute test scripts with limited human intervention. While these capabilities can accelerate testing activities, Rayal stresses that human review remains indispensable at the conclusion of the testing process. Validation professionals must evaluate AI-generated results, verify that identified defects have been fully resolved, and ensure no issues have been overlooked before approving system release.
Another important consideration is inspection readiness. As regulators increasingly encounter AI-generated validation records, organizations must ensure that documentation remains transparent and traceable. Rayal notes that inspectors are less concerned with whether AI was used and more focused on whether the validation process remains understandable and controlled. Audit-ready AI-generated rationales should clearly document the inputs used, explain how conclusions were reached, and maintain traceability to requirements, risks, testing activities, and approvals. Equally important is clear evidence that qualified personnel reviewed and approved all critical decisions.
Rayal identifies several decision points that should always require human sign-off, regardless of how accurate or confident AI recommendations may appear. These include determinations affecting patient safety, product quality, regulatory compliance, and data integrity. Specifically, intended-use definitions, system classifications, risk assessments, residual risk acceptance, deviation disposition, validation conclusions, and production release decisions should remain under direct human control. AI may provide insights and recommendations, but responsibility for these decisions cannot be delegated to algorithms.
For organizations considering AI adoption, Rayal presents a straightforward business case. Rather than replacing validation professionals, AI can free them from repetitive documentation and administrative tasks, allowing them to devote more time to expert analysis and patient-focused decision-making. When implemented within a strong governance framework that includes robust oversight and accountability, AI can improve consistency, increase efficiency, enhance inspection readiness, and help organizations manage growing workloads without compromising compliance or quality standards.
As discussions at the 2026 ISPE Annual Meeting & Expo are expected to highlight, the future of validation is not about choosing between artificial intelligence and human expertise. Instead, success will depend on determining when AI can be trusted to accelerate routine work and when human professionals must step in to verify, approve, and ultimately take responsibility for decisions that affect patients, products, and regulatory compliance.
Learn More About the 2026 ISPE Annual Meeting & Expo
About Pharmaceutical Online
Pharmaceutical Online is part of the Life Science Connect media group and serves pharmaceutical manufacturing, quality, engineering, and supply chain professionals with technical insights, industry news, best practices, and expert analysis. The publication covers topics including drug manufacturing, facility design, process development, automation, validation, regulatory compliance, packaging, and supply chain optimization, helping industry stakeholders stay informed about emerging technologies and operational strategies across the pharmaceutical sector.