Artificial Intelligence (AI) Is Changing Our Tools, Not Our Principles
One of the most anticipated sessions at the 2026 ISPE AI in Life Sciences Summit, Powered by GAMP® was a panel titled, “AI in Pharmaceutical Manufacturing is Outpacing Regulation.” Judging by the attendance, the title struck a chord. The room was full well before the discussion began, and when the session concluded there were still attendees waiting at the microphones and unanswered questions remaining in the online queue. It was clear this conversation had touched something much larger than a single conference topic.
AI has moved beyond curiosity. It is becoming part of the daily conversation in pharmaceutical research, manufacturing, quality, and regulatory affairs. Some organizations are already deploying sophisticated AI applications. Others are taking their first steps. Most of the industry is somewhere between those two points, learning where these technologies can create value while also understanding the responsibilities that come with using them.
I attended the session expecting to hear a debate about whether regulation is keeping pace with technology. Instead, I heard something much more encouraging.
The discussion was remarkably balanced. Industry leaders shared practical examples of AI improving operational efficiency, reducing deviations, accelerating analysis, and allowing experts to spend less time searching for information and more time solving meaningful problems. Regulators acknowledged both the opportunities and the challenges while repeatedly emphasizing collaboration. Throughout the discussion there was no sense of confrontation. Instead, industry leaders and regulators approached the discussion as partners trying to solve the same challenge. Both recognized that AI is advancing rapidly, both acknowledged there are questions still to answer, and both emphasized the importance of learning together.
Several themes emerged that I believe are worth reflecting on.
The first was that AI does not diminish the importance of expertise. In fact, it may increase it.
There is understandable concern that increasingly capable AI systems could replace human expertise. The discussion suggested something different. If AI assumes more of the routine work associated with collecting information, drafting documents, analyzing data, or identifying trends, experts become even more valuable because their time shifts toward interpretation, judgment, and decision-making. Those are precisely the activities that cannot simply be delegated to technology.
The second theme was that AI does not replace scientific principles.
Throughout the discussion, regulators consistently returned to concepts that pharmaceutical professionals already understand well: intended use, scientific justification, risk management, data integrity, and patient safety. AI introduces new capabilities, but it does not eliminate the need to understand how those capabilities are being used or to explain, with sound science, why they can be trusted.
One comment from a regulator captured this perfectly: “If you can't explain how you are using AI, why you trust it, and the science behind it, don't expect others—including regulators—to trust it either.” That struck me because it wasn't really about AI. It was about scientific credibility, something our industry has always valued.
The third theme was that trust is becoming more important, not less.
One of the audience questions addressed the growing use of AI by suppliers and contract manufacturers: “How do organizations manage AI governance if they do not always know where AI is being applied?” Rather than immediately discussing contractual language, the conversation turned toward trusted partnerships, transparency, and risk-based oversight. As AI becomes embedded within more business processes, relationships built on openness and trust become increasingly valuable. That observation extends well beyond supplier management. Trust underpins relationships between industry and regulators, technology providers and customers, and ultimately pharmaceutical companies and the patients they serve.
Another point that surfaced repeatedly was the importance of data.
Questions about knowledge graphs, ontologies, model selection, and governance all converged on one conclusion: AI can only perform as well as the data that supports it. Sophisticated models cannot compensate for incomplete, poorly governed, or inconsistent information. Organizations eager to accelerate their AI journey may discover that one of the most important investments they can make is not in a new model, but in strengthening the quality, accessibility, and governance of their data.
Perhaps the most encouraging aspect of the discussion was the willingness to acknowledge that none of us has all the answers.
The concept of "human in the loop" came up repeatedly, yet even regulators acknowledged that its practical implementation is still evolving. Panelists spoke about creating safe environments where industry and regulators can explore new ideas together, share experiences, and develop practical approaches before expectations become formal guidance. That spirit of collaboration may prove to be one of the industry's greatest strengths as AI continues to mature.
As I left the session, one thought stayed with me. For a panel focused on AI, the conversation was remarkably human. It centered on judgment rather than automation, scientific thinking rather than algorithms, collaboration rather than conflict, and trust rather than technology. Most importantly, every discussion eventually came back to patients. That should give all of us confidence that, as our industry embraces powerful new technologies, our purpose and our principles remain unchanged.
AI will undoubtedly become part of research and development, regulatory submissions, manufacturing operations, quality systems, and countless other activities across our industry. I believe these tools will help us analyze more information, solve increasingly complex problems, and remove much of the routine work that occupies talented scientists, engineers, and quality professionals today.
When that happens, the real opportunity will not simply be doing today's work faster. It will be freeing talented scientists, engineers, quality professionals, and manufacturing leaders to spend more of their time exercising judgment, solving difficult problems, asking better questions, and developing the innovations that improve the quality of human life.
The packed room reminded me that our industry recognizes both the opportunity and the responsibility before us. We are still early in this journey, but the conversation I witnessed gave me confidence that we are approaching it in the right way. By learning together, challenging one another, applying the scientific principles that have always guided our industry, and keeping patients at the center of every decision.
