Artificial intelligence is no longer a promise of the future – it has already become part of the systems that run our cities. It drafts emails, designs buildings, steers traffic, and powers civic chatbots. It answers before we finish asking. Sometimes it lies, sometimes it hallucinates – and yet it always responds. This issue of topos does not celebrate AI, nor does it dismiss it. Instead, it takes the machine seriously. For planners, architects and urbanists, AI is now embedded in the very fabric of urban life – silently shaping infrastructures, simulations, and decisions. But what values are hidden in its code? Who defines its assumptions? And what happens when the model becomes more legible than the street itself? We explore the promises and perils of a technology that reproduces the past more than it imagines the future, that optimises for efficiency but risks narrowing democracy. From hackers and ethicists to urban designers and policy-makers, this issue uncovers the ethical, political and practical questions behind AI’s urban influence. AI Pioneers asks not what AI can do for cities, but what it means when urbanism itself becomes a function of AI.

Between lies, hallucinations and seduction

It speaks fluently now – this thing we still call artificial intelligence, though the term has long since outgrown its own definition. It drafts our emails, draws our buildings, answers our questions before we’ve finished asking them. Sometimes it lies. Sometimes it hallucinates. Always, it responds. And that, of course, is the seduction.

Embedded in everything

This issue of topos does not celebrate AI. Nor does it warn you off. It does something more dangerous: it takes the machine seriously. For the urbanist, the architect, the planner, the administrator of increasingly digitalised complexity, AI is no longer a speculative force. It is embedded – silently and exponentially – in everything from traffic modulation to zoning simulation, from energy grids to civic chatbots. The city, once a physical terrain of political decision and social negotiation, is becoming something else: a system increasingly shaped by systems. But who built those systems? What values were trained into them? What assumptions lie dormant in the code? And – perhaps most urgently – what happens when the simulation becomes more legible than the street itself?

Not wisdom, but automation

AI, as deployed in today’s urban contexts, is neither neutral nor magical. It is a mirror polished by scale. It reproduces the visible past with stunning precision – and with it, the invisible exclusions. It recommends what is common, not what is just. It optimises for what already exists, not what ought to be. And in doing so, it threatens to turn the practice of urbanism into an act of retroactive confirmation. Cities, we are told, will be more efficient, more intelligent, more predictable. But one might also say: more predictable is less democratic. More optimised is less open. And more intelligent – if that intelligence is detached from judgement – is not wisdom, but automation.

Who is able to think?

In these pages, you will find no techno-utopian platitudes. You will find hackers who know what it means to switch off a city. You will find ethicists who question the morality of training machines on stolen text. You will read about model bias, infrastructural fragility, and the quiet geopolitical centralisation of cognitive power. You will see cities rendered by code – and cities distorted by it. And if we are lucky, you will begin to ask better questions. Not “what can AI do for urbanism?”, but: “what does it mean when urbanism becomes a function of AI?” Because in the end, the great urban question is not one of tools, but of control. Who gets to plan? Who gets to decide? And who – when the lights go out, when the voice stops answering, when the model fails – is still able to think? This is not an issue about intelligence. It is an issue about responsibility.

I hope you read it with curiosity, with discomfort – and above all, with your mind firmly switched on. With this in mind, I look forward to your feedback, suggestions and all the discussions that will be sparked by this issue.

Get the topos 132– AI pioneer – here.

Our last issue is all about the 15-minute city. The concept of the 15-minute city is causing a stir in urban planning worldwide. It promises liveable, walkable neighbourhoods in which everything important – from schools and healthcare to workplaces – can be reached within 15 minutes. Paris, Barcelona and Melbourne have already adopted this model. But can it really improve everyday life, reduce emissions and strengthen the sense of community? In view of the World Bank’s forecast that almost 70 per cent of humanity will live in cities by 2050, the question arises: is the 15-minute city more than just an ideal? We shed light on the potential, limits and concrete realisation of a concept that could fundamentally change our cities. Read more in the editorial of topos 131 – 15 minute city.