There really won't be anything that AI can't do better than humans, apart from being human, perhaps.
— Elon Musk, speaking to The Economist's Zanny Minton Beddoes, July 2026
Most people read that as a warning. I find it reassuring.
What the Profession Is Actually For
The value of community engagement has never really been in producing documents.
It is in building trust. Reading the room. Understanding concerns that people find difficult to articulate — sometimes because the concern is technical, more often because it is personal. It is in helping communities and decision-makers work through choices where there is no option that leaves everyone satisfied.
That work is not a communications exercise. It is judgement applied under scrutiny, in public, usually with a history behind it.
None of it is document production. Documents are the artefact left behind.
What AI Is Genuinely Good At
AI is already very capable at drafting, summarising, analysing and organising information. Those capabilities are improving quickly, and that is a good thing.
Consider where practitioner time actually goes on a typical project. Building the engagement plan. Structuring the stakeholder analysis. Reading several hundred submissions and grouping them into themes. Writing the What We Heard report. Preparing the council paper. Reformatting the same content for four different audiences.
Much of that is structured, repetitive and rules-based. It is exactly the kind of work that AI handles well — and exactly the kind of work that expands to fill whatever time is available.
If AI takes more of that off our hands, practitioners get time back for the part of the job that has always mattered most: the conversations, the relationships and the judgement that define good engagement.
The Part That Doesn't Transfer
There is a reason the qualifier in Musk's line matters.
Being human is not a residual category of leftover tasks. In engagement, it is the substance of the work.
A model can summarise what a community said. It cannot sit in a hall while someone describes what a proposed road will do to the street they have lived on for forty years, and register what that means for how the process should proceed. It can identify that sentiment is negative. It cannot tell you whether the negativity reflects a design problem, a trust problem inherited from the last project, or a failure to explain a constraint properly.
It can produce a defensible-looking synthesis. It cannot be accountable for it. Someone has to stand behind the conclusion when it is challenged, and that someone is a practitioner.
This is the line we have written about before — the difference between AI supporting judgement and substituting for it. Time saved on preparation only helps if that line holds.
The Risk in This Framing
It would be easy to end there, and it would be too neat.
Freed capacity does not automatically become engagement capacity. It can just as easily be absorbed — more projects per practitioner, faster turnarounds, the same headcount asked to cover more ground. The efficiency is real, but nothing about it guarantees the time returns to the community.
That is an organisational choice, not a technology outcome. If AI halves the time it takes to produce an engagement plan, someone decides whether the saved hours go into more site visits and harder conversations, or into a larger project portfolio.
The second choice is the more likely default. It is also the one that quietly erodes the thing the efficiency was supposed to protect.
What This Means for Practitioners
Two things follow.
The first is that the skills worth investing in are the ones that do not transfer. Facilitation under pressure. Building relationships with people who have reason to distrust the process. Explaining a constraint honestly rather than defensively. Knowing when a technically sound recommendation is politically or socially untenable, and saying so. These have always been the hardest parts of the job. They are about to become the most distinctive.
The second is that the case for time returned to the community has to be made deliberately, and made to the people who set workloads. It will not happen on its own.
Perhaps that is what the future looks like.
Less time preparing engagement.
More time engaging.
Related reading: what responsible AI looks like in engagement practice, the risks of generic AI in community engagement, and how to use AI in community engagement planning. For the full overview, see AI in community engagement: what actually matters.