It’s 9am. You have a brief that says “make it engaging,” a 200-page SME document, and a blank slide deck. You know this moment. AI can now handle it. Synthesise the document, draft the objectives, sketch the outline, write the learning outcomes using Bloom’s taxonomy. In an hour, not a day.
So, what does that actually mean for the people who used to spend the day doing exactly that?
Thoughts of an ID: There is a particular kind of professional who has spent years becoming very good at tasks that were never, strictly speaking, the point. One sees them in every field: the lawyer who can cite precedent from memory but struggles to advise a client; the doctor whose diagnoses are immaculate but whose patients leave more confused than when they arrived.
They are not incompetent. Far from it. They have simply mastered the scaffolding and, in the process, grown so comfortable there that the building itself has become almost beside the point.
One does not blame them. The scaffolding was always what was asked for.
What shifted
The tasks AI handles well are real and significant: first-draft content generation, learning outlines, source material synthesis, objective writing. These haven’t disappeared, they’ve been compressed. A day’s work now takes an hour.
That compression matters because those tasks were always preparation for the real work, not the work itself.
What the real work looks like
The judgment calls AI cannot make: Is the brief solving the right problem? Is training even the right intervention? What does this organisation’s context mean for how this design needs to behave?
An AI can produce a well-structured five-module course on conflict resolution in 90 seconds. It cannot tell you that your client’s teams avoid direct feedback not because they lack the skill, but because the last person who gave it was managed out. That distinction determines whether the course works or sits on an LMS untouched.
Thoughts of an ID: Human beings are, on the whole, considerably more interesting than the problems attributed to them. I have observed, over the course of many years, that when an organisation commissions a training programme on communication, it rarely suffers from a deficit of information about how to communicate.
It suffers from something older and more stubborn: a culture that punishes honesty, or a history that has made trust expensive, or a leadership that says one thing on a Monday and demonstrably does another by Friday.
No course corrects these things. But a perceptive designer, sitting in the right room, asking the right questions with genuine curiosity rather than professional obligation, might at least ensure that the course does not make them worse.
The skills gaining value
- Critical evaluation: not whether the output reads well, but whether it’s pedagogically sound, contextually accurate, and safe to deploy.
- Learning strategy: thinking upstream of content, questioning whether a training solution is actually what the performance gap requires.
- Systems thinking: understanding how a learning solution fits into the broader performance environment.
- Stakeholder fluency: translating between business need and design decision. Half diplomacy, half diagnosis.
Prompt writing is not on this list. It’s a useful skill, not a competitive edge. The edge is knowing what to do with the output… and when to discard it.
The risk worth naming
Speed is seductive. AI-generated content arrives polished, and “good enough” is easier to defend when it looks finished. The pressure to take the first draft, tweak it lightly, and ship it, without ever interrogating whether it addresses the real problem, is significant.
Thoughts of an ID: I once knew a man who prided himself on never wasting time. He was extraordinarily efficient. He answered correspondence within the hour, produced reports of admirable brevity, and had the curious ability to walk out of a meeting before anyone had quite noticed he had stopped contributing to it. His colleagues admired him greatly.
It was only much later, when the work he had done so efficiently began to unravel in the field, that anyone thought to ask what, precisely, he had been so busy not wasting time on. Speed, I have come to think, is a virtue in the same way that tidiness is a virtue: useful when in service of something, and a mild pathology when mistaken for an end in itself.
The risk is not that AI replaces instructional designers. It’s that some instructional designers let it replace their thinking.
An overdue shift
The field has long argued it gets reduced to content production. Stakeholders treating IDs as execution resources. Projects where the brief arrives and thinking is not part of the contract.
AI is creating genuine pressure to move upstream. The consultative conversations, the strategic decisions, the design choices that actually change what people do at work. That space has always been where the field wanted to be.
Thoughts of an ID: There is something almost poignant about a profession that has spent decades arguing for a seat at the table, only to find, when the chair is finally offered, that it requires sitting in a different room entirely.
The table has moved. The conversations happening there are less about content and more about problems, less about outputs and more about outcomes. Whether one finds this exciting or alarming probably says something about the kind of designer one has chosen to become.
The question is whether the field steps into that space, or uses faster tools to do more of the same.
Has AI changed how you spend your time as an ID? Has it created room for better work, or just pressure to produce more of it?