MentherisWhat happens when AI removesthe path to becoming an expert?
Automation may change more than the work we do. It may change how expertise is built.
Most conversations about AI and work begin with a familiar question: which tasks can be automated? But another question sits underneath it: what were people learning while doing those tasks?
01 · Looking beyond the task
Automation evaluates outputs.Expertise develops through experience.
Expertise rarely appears fully formed. In many professions, it develops gradually: through routine work, repeated exposure, mistakes, judgement, increasingly difficult decisions and accumulated experience.
Junior work therefore has more than one function. It produces an immediate output, but it can also create the experience from which future capability develops.
A task can produce both an output and a learner.
We automate tasks.But careers are built through sequences of tasks.
02 · Reframing automation
Perhaps the unit of analysis is too small.
A task may look replaceable when considered in isolation. But careers and expertise develop across sequences of tasks, interactions and increasing responsibility. Evaluating automation one task at a time may therefore miss what that task contributes to a longer developmental pathway.
This does not mean routine work should be preserved simply because it has always existed. It means that before removing a task, we may need to understand what people were becoming capable of by doing it.
03 · Emerging questions
What do we need to understand before the pathway changes?
Which tasks carry developmental value beyond their immediate output?
What forms of experience remain necessary for developing judgement?
How can new learning pathways reproduce capabilities that previously developed through junior work?
Who becomes capable of challenging an AI system if fewer people develop deep expertise in the systems AI increasingly mediates?
This note does not argue that expertise pipelines are already collapsing across professions. The observation is narrower: AI-driven automation may alter not only the distribution of work, but also the mechanisms through which expertise is reproduced.
That relationship deserves closer investigation.
The question is not only what AI can replace.It is what must still be learned.
This Research Note develops an emerging question raised during the first Mentheris Research Dialogue in May 2026, where participants discussed how AI adoption may affect established expertise pathways in technically sophisticated fields.
Continue the inquiry
Are you seeing this happen in your field?
If AI is changing junior work, training pathways or the way expertise develops in your profession or organisation, I would be interested in understanding what you are observing.