When the profession outruns the mentor

AI is redefining work faster than experience can be passed on.

Mentorship has traditionally depended on mentors having experience that mentees have not yet had the opportunity to acquire. But what happens when the profession changes faster than experience can accumulate?

I’ve spent most of my professional life on both sides of mentorship. Early in my career, I benefited from mentors who patiently helped me develop as a designer. As my career progressed into design leadership, I hired junior designers, reviewed portfolios, coached employees through difficult projects, and watched inexperienced designers grow into confident professionals. Today, as a professor, I mentor students entering a profession that’s changing at an unprecedented pace, while continuing to benefit from mentors in my own academic career.

For most professions, that arrangement has worked remarkably well. Experience compounds over time because the profession itself evolves gradually. Mentors don’t simply know more—they know things that remain relevant long enough to pass on to the next generation. Mentees benefit because they’re preparing for essentially the same profession, only a few years later.

That dynamic has held together everything from apprenticeships to graduate school. A carpenter teaches another carpenter because the fundamentals of building remain relatively stable. Physicians mentor residents because human anatomy isn’t rewritten every product cycle. Even technology, despite moving faster than almost every other industry, generally changed slowly enough that experience continued to accumulate faster than the profession evolved.

When I began my career, web design looked very different from what it does today. I remember building layouts using HTML tables before CSS became the standard, watching Flash rise and disappear, adapting to responsive design, and eventually finding myself designing interfaces while also writing the front-end code behind them. The tools changed constantly, but the profession itself remained recognizable and predictable enough that someone with years of experience could confidently mentor someone just entering the field.

Artificial intelligence is different because it isn’t simply replacing one tool with another. It is beginning to redefine the work itself, creating a kind of disruption unlike anything most professions have previously experienced. Replacing Photoshop with Figma is one kind of change. Redefining what it means to be a designer is something altogether different.

When I entered the design field, becoming competent followed a fairly predictable path. You learned typography by making typographic mistakes, developed an eye for composition through critique after critique, and gradually built the judgment to know why one solution was stronger than another. Over time, you weren’t just learning software — you were developing taste, visual intuition, and the ability to make design decisions with confidence.

Today, a student or junior designer can generate interface concepts, illustrations, icons, personas, user flows, and even usability recommendations in a matter of seconds. Whether those outputs are exceptional or merely good enough is almost beside the point. What matters is that this changes the traditional pathway through which designers have historically developed expertise.

The more I think about it, the more I believe mentorship has always relied on an assumption we’ve largely taken for granted: experience remains valuable because professions typically evolve more slowly than careers.

Artificial intelligence introduces a very different dynamic. Portions of professional expertise now seem to have something resembling a half-life, depreciating not because experienced professionals suddenly become less capable, but because the profession itself is evolving while careers are still unfolding. A designer with twenty years of experience may have only two or three years of experience working in an AI-first environment.

That observation isn’t meant as a criticism of experienced professionals. It’s simply a recognition of how rapidly the profession is changing. Many mentors and mentees are now learning portions of the profession simultaneously, creating a relationship that looks less like traditional apprenticeship and more like collaborative exploration. That is a profound shift in what mentorship itself means, and I don’t think we’ve fully reckoned with its implications.

The philosopher Michael Polanyi argued that “we know more than we can tell,” describing what he called tacit knowledge — the intuition and judgment that emerge through experience rather than formal instruction. His work became enormously influential because it explained why mentorship matters in the first place. A mentor doesn’t simply transfer information—they transfer ways of seeing that cannot easily be written down or reduced to a checklist.

Large language models complicate that picture in fascinating ways. They can often produce work that resembles the output of an experienced professional without possessing the lived experience that traditionally produced it. The result is that they imitate expertise without necessarily acquiring it, allowing beginners to generate work that previously required years of proficiency.

One might assume that if experience is no longer required to produce work that appears professional, then expertise no longer matters. I think that’s exactly the wrong conclusion. If anything, expertise becomes more valuable because generating an answer and recognizing a good answer are fundamentally different skills.

That second skill isn’t something AI can hand you — recognizing quality still requires the calibration that only comes from having struggled to produce it yourself, made the wrong call, and learned what better actually looks like.

That realization should also change what mentors spend their time teaching. While mentorship has always involved cultivating judgment, much of the focus has traditionally been on execution — how to write better code, compose stronger layouts, conduct better research, or manage increasingly complex projects.

As AI assumes more of that execution, I think the emphasis shifts toward helping mentees evaluate AI output, recognize weak reasoning, identify hidden assumptions, and understand why one solution is genuinely better than another.

The same shift applies to education. I no longer think my primary responsibility is helping students master today’s software because today’s software is unlikely to remain today’s software for very long. My responsibility is helping students develop the habits of mind that allow them to survive whatever replaces it.

Mentees should probably rethink what they look for as well. Twenty years ago, I would have sought the person with the deepest technical expertise and the longest resume. Today, I’d be looking for someone who has successfully adapted through multiple technological revolutions without becoming either dogmatic or infatuated with every new trend that promises to reinvent the world.

That’s a surprisingly high bar because adaptability is much harder to imitate than expertise. It isn’t demonstrated by what someone already knows, but by how they respond when what they know is no longer enough. It requires intellectual humility, curiosity, and the willingness to admit that yesterday’s success doesn’t automatically explain tomorrow’s profession.

Perhaps AI isn’t exposing a weakness in mentorship so much as revealing what it was always supposed to be. The best mentors were never valuable for knowing the most. They were valuable for being able to explain why something worked, not just that it did.

That’s about to become the whole job. A model can generate six interface concepts before a junior designer finishes a sketch. What it can’t do is tell you why option four’s hierarchy fails or why the spacing fights the content. That explanation is the mentorship now, and it’s the one part AI didn’t supply.

Mentors who keep teaching execution, how to produce the work, will find that’s exactly the part AI now does on its own. Mentors who teach someone to say precisely why a solution is weak are teaching something a faster model can’t make obsolete.

Don’t miss out! Join my email list and receive the latest content.

When the profession outruns the mentor was originally published in UX Collective on Medium, where people are continuing the conversation by highlighting and responding to this story.