Cape Town's AI ambition.
The questions nobody is asking.

On August 11, 2026, Mayor Geordin Hill-Lewis spoke at the BizNews Conference. What he said was specific, not aspirational. He would not grant departmental managers additional budget for salaries unless they could demonstrate how they were introducing AI adoption and automation in their department. Not eventually. Not in principle. As a condition of the salary request being considered at all.

Hill-Lewis said he anticipated leadership adoption of the tech would be the biggest hurdle to overcome, and this was how he was addressing it. He had become increasingly reluctant to approve requests that would simply expand the city's bureaucracy. The mechanism he chose to drive adoption — tying salary budget access to AI implementation plans — is not a gentle nudge. It is a hard requirement.

This is not happening in isolation. The City of Cape Town has launched Maya, an AI-powered investment concierge available 24 hours a day in eight languages including Afrikaans and isiXhosa — designed to help global investors navigate municipal approvals, identify commercial opportunities, and connect with City officials. AI traffic cameras have been piloted on Philip Kgosana Drive, detecting seatbelt violations and phone use simultaneously. The City has approached the National Director of Public Prosecutions for guidance on legal admissibility before a wider rollout.

South Africa has been identified as Africa's most AI-ready economy, ranking 46th out of 147 economies in Microsoft's 2026 Global AI Diffusion Report. Cape Town wants to be the continent's AI hub. The vision is coherent and, in some ways, genuinely exciting. The question is what it means in practice for the tens of thousands of people who currently work for the City.

What happened in the United States — a cautionary reference point

The United States ran the most dramatic recent experiment in government AI adoption and workforce reduction, and the results are instructive — not as a prediction of what will happen in Cape Town, but as a documented case study of what can happen when the pressure to adopt AI and the pressure to reduce headcount arrive at the same time.

Through 2025, DOGE initiated mandatory attrition targets, reductions-in-force plans, a hiring freeze, and contract cancellations that removed contractor employees. The actions triggered 348,219 individuals to quit, retire, be laid off, or otherwise leave federal employment. The stated purpose was efficiency and cost reduction. AI was both the justification and the replacement mechanism.

After mass firings at the General Services Administration, DOGE issued remaining staff an app to take up the excess work — but severely limited its use to a handful of repetitive tasks. Fears of AI-driven mass firings of federal workers were described as not unfounded, with automated reduction-in-force software potentially involved in layoff decisions.

The outcome was not a leaner, more efficient government. Following massive workforce reductions and a $165.6 billion hit to the US economy, federal managers were struggling to integrate AI as low engagement collapsed across agencies. The people who knew how to do the work were gone before the AI that was meant to replace them actually worked.

Cape Town is not the United States federal government. Hill-Lewis is not Musk. The political context, the labour law environment, and the stated intention are all different. But the dynamic — leadership pressure to adopt AI as a precondition for budget approval — is structurally similar. And the question it raises is the same: if AI adoption is the condition for salary budget approval, what happens when AI adoption reduces the number of salaries the department needs?

The seven questions the City of Cape Town should be asked about AI adoption

These are not rhetorical questions. They are genuine uncertainties that the Mayor's AI push creates for City employees, Cape Town residents, and anyone whose career is built around municipal employment. They deserve direct answers — and if the answers are not yet available, they deserve to at least be asked publicly.

What a genuinely AI-advanced city looks like

The vision Hill-Lewis is articulating — a Cape Town that is genuinely technologically advanced, that uses AI to improve service delivery, attract investment, and operate more efficiently — is worth building. There is nothing wrong with the ambition. The cities that will matter most in the next decade will be the ones that got AI adoption right in their public institutions, not the ones that were cautious.

But there is a version of this that works and a version that does not. The version that works uses AI to genuinely improve services for residents at every income level. It is transparent about workforce implications and builds retraining and transition support before those implications materialise. It verifies actual impact rather than accepting adoption plans at face value. It deploys technology that serves the people who live in the city, not just the people who want to invest in it.

The version that does not work uses AI adoption as a budget mechanism — a way to hold salary increases hostage to technology adoption without a corresponding commitment to the people whose work is being transformed. It produces impressive-sounding announcements without changing what it is like to apply for housing, report a burst pipe, or appeal a rates assessment. It displaces workers faster than it creates alternative pathways for them.

Cape Town has the talent, the investment appetite, and the political will to build the version that works. The question is whether the people leading the AI push are asking themselves the hard questions clearly enough to avoid the version that does not.