The Expertise Question that Artificial Intelligence (AI) Cannot Answer

A team of pharmaceutical professionals is facing a manufacturing challenge that threatens the supply of an important medicine. The situation is unusual. It does not seem to fit what is known. Previous investigations have not uncovered the root cause. The team uses a highly advanced AI system. Within seconds, it searches decades of regulations, technical papers, guidance documents, operational records, and industry knowledge. The system identifies similar situations. It proposes possible explanations. It suggests several paths forward.

Yet the problem remains unsolved.

Not because the AI failed, but because the answer does not exist in the knowledge it was trained on.

No one has encountered this exact problem before. No one has conducted the experiments. No one has challenged the assumptions. No one has generated the expertise required to create a new answer.

At that moment, the industry needs something AI cannot provide on its own.

It needs experts.

I recently had a fascinating conversation with a new colleague from NUM Solution about AI and its growing role in our professional lives. During that discussion, we arrived at a question that has stayed with me:

If future generations increasingly learn from AI, who will create the expertise that future AI systems depend upon?

The question is particularly relevant because AI is already delivering tremendous benefits.

Today's professionals can access information at a speed that would have been unimaginable only a few years ago. A new engineer can quickly understand a technical concept. A project team can identify relevant guidance in minutes rather than hours. A professional can explore unfamiliar topics without spending days searching through documents.

The productivity gains are remarkable.

Knowledge that once took years to accumulate can now be accessed almost instantly. Professionals can become effective contributors much faster than previous generations. Organizations can make decisions more quickly. Teams can spend less time searching for information and more time applying it.

These are significant advantages, and we should fully embrace them. Personally, I wish I would have had access to this kind of knowledge when I entered the industry.

Yet there is an important distinction between accessing knowledge and creating expertise.

Expertise is not simply information. Expertise is information tested against reality.

Anyone who has spent time in pharmaceutical manufacturing understands this difference. Some of the most valuable knowledge in our industry cannot be fully captured in a document or a database. It is developed during technology transfers, startup activities, inspections, investigations, deviations, and unexpected equipment behavior. It is developed through experience.

Knowledge resides with:

• The experienced operator who notices something unusual before an alarm activates
• The engineer who recognizes a risk because they have seen a similar situation before
• The quality professional who understands that a technically correct answer may not be the best answer
• The project leader who knows when a proven approach should be challenged because circumstances have changed

That kind of judgment cannot simply be retrieved. It must be developed.

This raises an important concern.

If professionals increasingly rely on AI to provide answers, will enough people still undertake the journey required to become true experts themselves? Will they gain the experience necessary to create the next generation of industry knowledge?

It is a fair question.

But history suggests that technological advancement does not eliminate expertise. Instead, it changes the nature of expertise. The printing press did not eliminate scholars. Calculators did not eliminate mathematicians. Computer-aided design did not eliminate engineers. And the internet did not eliminate researchers.

Each of these innovations made information more accessible and routine tasks more efficient. Yet the demand for deep expertise did not disappear. In many cases, it became even more valuable because experts could focus less on gathering information and more on solving complex problems.

Perhaps AI will follow the same path.

By reducing the effort required to find and organize knowledge, AI could free experts to spend more time experimenting, innovating, collaborating, and developing entirely new knowledge.

If that happens, expertise generation may accelerate rather than decline. The outcome will depend on how we choose to use these tools.

This is where organizations like ISPE have an important role to play.

Historically, professional societies have captured expertise, validated expertise, published expertise, and transferred expertise to new generations of professionals. In an AI-enabled future, our most important contribution may increasingly be the creation and validation of knowledge itself.

Consider how much of the industry's progress has emerged. It has not come from a single company or a single expert working in isolation. It has come from practitioners sharing experiences, debating ideas, challenging assumptions, conducting research, and collectively determining what works.

That process remains essential.

When ISPE members contribute to guidance documents, participate in Communities of Practice, present lessons learned at conferences, engage in training programs, mentor emerging professionals, and collaborate across organizations, they are doing more than sharing information.

They are contributing to the future body of knowledge for our industry. They are transforming individual experience into collective expertise. They are helping ensure that the next generation of professionals inherits not only answers, but also the ability to discover new ones.

Perhaps this is where professional societies become even more important in the age of AI.

• Professional communities create knowledge.
• AI can distribute it.
• Professional communities help us tackle questions that have never been answered before.
• AI can help us find existing answers faster than ever before.
• Professional communities develop deeper questions and explore greater possibilities to fuel further innovation.

The future of our industry will depend on both.

As AI continues to advance, we should embrace its ability to accelerate learning, improve productivity, and broaden access to information. But we should also recognize that every AI system ultimately depends on knowledge that was first created by people willing to explore the unknown.

The future will still need experts. It will still require experimentation. It will depend on practitioners willing to challenge assumptions and solve problems that have never been solved before. And perhaps most importantly, it will rely on communities that bring those people together.

ISPE’s Body of Knowledge is one of its greatest strengths, built by thousands of members who, over decades, have taken the time to share what they’ve learned through experience. Their contributions have created a strong and trusted foundation for the industry.

We are working to capture and organize this knowledge in a more structured way, leveraging the new capabilities of AI to develop an AI-driven knowledge tool for the life sciences industry, to make it easier for practitioners, from recent graduates to seasoned professionals taking on new challenges, to access, share, apply insights from ISPE’s Body of Knowledge.

That is a responsibility worth embracing. It is also an opportunity for organizations like ISPE to help ensure that expertise continues not only to be shared, but also to be created for generations to come.

Join and explore opportunities to contribute to the ISPE community.