AI Doesn't Replace Clinical Informaticists. It Multiplies Our Impact.

It's understandable that many clinical informaticists are wondering what AI means for our profession.

Every week brings another demonstration of AI writing software, answering questions, summarizing clinical information, and automating work that once required specialized expertise. If your career has been built around improving healthcare through technology, it's reasonable to wonder whether AI will eventually replace that role.

I believe the opposite is true.

I believe AI will dramatically increase the demand for clinical informaticists because it removes one of the greatest constraints our profession has always faced: the difficulty of turning clinical insight into software.

Clinical informaticists occupy a unique intersection of medicine, technology, and systems thinking. We understand how care is actually delivered, where clinicians struggle, how workflows evolve, where safety risks emerge, and how technology can improve care without unintentionally making it worse.

For decades, our greatest limitation has rarely been a lack of worthwhile ideas. Instead, it has been our ability to transform those ideas into working systems.

We've all experienced some version of the same conversation:

"That's a great idea."

"It would absolutely help clinicians."

"But we don't have the engineering resources."

"Maybe next year."

Eventually, those conversations shape the way we think. We stop asking, "What would improve patient care?" and start asking, "What can we realistically get built?" We become experts not only at identifying opportunities, but at deciding which worthwhile ideas will never be pursued because implementation is simply too expensive.

That scarcity didn't just shape our work, it obscured our true identity.

Historically, much of our work has been translation. We translate clinical workflows into requirements, requirements into technical specifications, and technical limitations back into language clinicians can understand. Translation has always been necessary—but it has also been one of the greatest bottlenecks to innovation.

Over the past several months, I've spent a great deal of time building software with AI. I don't mean generating snippets of code. I mean designing, implementing, debugging, and iterating on real systems.

I've spent forty years in technology, with twelve years of that in clinical informatics, and I'm also a registered nurse. In all that time, I've never experienced a technological shift quite like this one.

Using AI to build software genuinely feels like acquiring a superpower—not because it replaces my expertise, but because it allows me to express that expertise in ways that were previously impossible.

Ideas that I would have dismissed only months ago as impossibly expensive—or simply impossible—are now not only feasible, they're often surprisingly easy to prototype.

That experience fundamentally changed how I think about AI.

More importantly, it changed how I think about clinical informatics.

The most important thing AI is changing isn't software.

It's constraints.

For most of my career, software has been expensive to create. Every worthwhile idea had to compete for limited engineering capacity, limited budgets, and limited organizational attention. AI dramatically lowers those costs. It compresses the distance between an idea and a working system.

When one constraint disappears, something else becomes the constraint.

For years, implementation was the bottleneck.

Increasingly, clinical systems thinking will be.

Our value has never been in translating ideas.

Our value has always been in having the right ideas.

As implementation becomes dramatically less expensive, translation becomes less of a bottleneck. What becomes more valuable is clinical systems thinking— the ability to recognize meaningful problems, understand complex workflows, anticipate unintended consequences, and envision better ways to deliver care.

Code is becoming abundant.

Translation is becoming effortless.

Clinical systems thinking is not.

You are the scarce resource.

AI can generate thousands of lines of code. It cannot recognize a dangerous workflow, understand the lived experience of clinicians, balance the competing priorities of safety and usability, or determine what is worth building.

Those are exactly the decisions we have spent our careers learning to make.

That naturally raises an important question.

If software becomes dramatically easier to build, does the role of the clinical informaticist become less important—or more?

I believe the answer is more—not because AI creates new clinical problems, but because it finally allows us to address problems that have always existed.

Some industries eventually approach "good enough." The essential problems are solved, demand levels off, and additional investment produces diminishing returns.

Healthcare is different.

Healthcare has no practical finish line.

We will never decide that patient safety is good enough. We will never decide that clinician burnout has been solved. We will never stop wanting earlier diagnoses, safer workflows, lower cognitive burden, fewer adverse events, better patient experiences, or more equitable care.

Every problem we solve reveals another worth solving. Every minute we return to a clinician can be reinvested in caring for patients. And every improvement raises our expectations for what healthcare should become.

For decades, we have not been limited by the number of worthwhile ideas.

We have been limited by our ability to build them.

Economists have long observed a counterintuitive phenomenon known as the Jevons paradox: when technological advances make a valuable resource dramatically more efficient to use, total consumption of that resource often increases rather than decreases.

That's because lower costs don't simply replace existing work—they make entirely new work economically feasible.

I believe that's exactly what we're seeing in clinical informatics.

AI makes software implementation and translation dramatically less expensive. As those costs fall, organizations won't simply build today's projects faster. They'll pursue projects that were never even considered because they were too expensive to justify.

AI isn't creating demand for clinical informatics. It's revealing demand that has been there all along.

Organizations won't have to choose only the five highest-priority opportunities. They'll be able to pursue fifty.

That doesn't reduce the need for clinical informaticists.

It multiplies the impact of every clinical informaticist.

Ironically, the traditional translator role becomes less important while the profession itself becomes more important. Our value has never come from writing requirements. Our value has always come from understanding healthcare well enough to know what ought to exist. AI simply removes many of the barriers between that insight and the software that brings it to life.

For decades, we've become exceptionally good at prioritizing because we had no choice. We learned to say no to good ideas because implementation was expensive.

I hope the next era looks different.

I hope we spend less time asking, "Can we afford to build this?" and much more time asking, "What else could we build to make healthcare better?" That is a far more exciting question.

The age of AI doesn't lower the ceiling for clinical informaticists. It raises the ceiling on the impact we can have.

Our challenge is no longer simply keeping up with technology.

Our challenge is imagining—and building—a future worthy of the tools we now have.

Because for the first time in our profession, our ambition—not our engineering capacity—may become the limiting factor.