Do you remember when pathologists started finding each other on Twitter (now X) and #PathTwitter burst onto the scene? Some of us viewed social media as a distraction – so informal, unvetted, and outside the familiar channels of professional discourse. But others found something that was harder to name – a community formed in real time, without a blueprint.

Collaborations happened (and are happening), papers were written, and ideas traveled faster than any journal could carry them. Institutions and organizations caught up, and social media guidelines developed. I built an educational initiative called Pathodoodles on social media around the same time – in the space between the arrival of something new on our screens and the subsequent rules of engagement.

We are in a similar interval now. The difference is the pace.

Artificial intelligence (AI) entered medical education not through a coordinated rollout but through individual acts of curiosity, both from educators and students. The parallel I draw to social media is, honestly, imperfect – because social media asked us to reconsider how we communicated, while AI is asking us for something larger and more existential. It is asking us, rather forcefully, what our expertise even means when a tool can draft a lecture, generate a case, or produce a podcast from a slide deck in the time it takes to find a parking spot on campus. This is a different order of disruption, and it would be dishonest to minimize it.

Since last year, I have been facilitating faculty workshops on AI integration, contributing to institutional and professional workgroups, and developing a framework for how educators might approach these tools with intention and mindfulness. Here’s what that work has taught me: the feeling of trying to work with a moving target does not go away with experimentation and experience.

The pressure to stay current is real, the landscape shifts before our guidelines catch up, and there is no established expert you can model yourself on because anyone and everyone credible enough to teach this is also still learning it. What changes is not certainty, but our tolerance for productive uncertainty and our ability to act usefully within it.

A recent article in The Pathologist made the case for what AI changes on the curriculum side – shifting pathology education away from memorization, and toward judgment, interpretation, and reasoning through uncertainty. But that model gives rise to a new question: if judgment is what we are now asking of our students, what does it ask of the people teaching it, while we are still developing the judgment ourselves?

To answer this, I want to remind you of the habit we all have as pathologists. Before reaching for the microscope, we always learn about the clinical details and the gross specimen: size, shape, margins, cut surfaces. Only then do we go deeper with high-power microscopy. That sequence built within us isn’t about caution – it’s what makes that cellular detail meaningful.

AI works the same way. The educators getting the most from AI tools today are the ones who orient themselves first – defining the objective, setting the constraints, and gathering the source material before asking AI to build anything. The macroscopic context is what makes that microscopic detail so useful. This realization didn’t come to me from just reading about AI – it came from building a tumor nomenclature exercise.

As a pathology educator, I have a recurring problem with new learners constantly struggling with naming of neoplasms. Yet, the logic behind why something is called an adenoma vs carcinoma, and what that distinction tells you about how a tumor will behave, are basic steps to building a deeper knowledge about oncology. After a short training at Harvard Macy Institute that converted me from a skeptic to a builder, I was eager to experiment with AI to help tackle this problem.

I started with the goal and not the tool – students need to be able to apply naming logic, not just recall it. From there, I identified my own notes and a review article as sources, decided on an interactive format, and then picked my tools. NotebookLM helped turn my sources into a study guide and a short podcast, and Claude generated the HTML code for an interactive exercise, hosted as a GitHub page.

I want you to bear in mind that AI only did the drafting and formatting – what it couldn’t do was decide whether the classification logic it generated was precise enough to build correct clinical reasoning. That decision – the sign-out, so to speak – stayed with me. A board-certified pathologist confirmed it. Medical students used it. A second-year medical student wrote to me saying, “I am creating names now. Papilloma finally makes sense.” It’s a small thing. It is also, I think, the whole point.

This is, then, probably the answer to the question I asked earlier. If judgment is what we’re asking of our students, part of what it asks of us is to be visibly engaged in that judgment ourselves. We should approach these tools with the same mix of curiosity and caution we hope they’ll develop. Wonder and guardrails must stand together rather than be traded off. Rather than performing an expertise we don’t have, we should be modelling the orientation itself: pausing before prompting, being explicit about what we’re trying to achieve, and willing to say when something didn’t work.

The social media moment in pathology produced something lasting – a global community. New forms of scholarly communication were introduced, along with a professional infrastructure that no-one had planned, but everyone recognized as valuable. I do not know, or claim to know, what this AI moment will produce. What I do know is that the educators shaping how this technology lands in medical education are the ones engaging with it now – orienting first, building deliberately, and being honest about what worked and what didn’t.

The window between arrival and codification isn’t a gap to be anxious about. It is an opportunity, if we use it well. We have been here before. We know what to do.

Newsletters

Receive the latest pathologist news, personalities, education, and career development – weekly to your inbox.

I have read and understand the Privacy Notice*

Subscribe