Source: Unsplash
Cities around the world are installing AI-powered traffic systems, building digital twins of entire neighbourhoods, deploying sensors on every street corner. But walk through any of these "smart" cities as an actual resident and ask yourself: who designed this for me?
Smart cities are suffering from the same design failure that plagued enterprise software for decades: they are built from the top down by engineers, governments, and vendors, with the actual users as an afterthought. The UN held its 13th World Urban Forum in Baku just last month, and the central concern raised by experts was not whether the technology works. It is whether anyone is designing it for the people who have to live inside it. The answer, in most cities, is no.
"Trust today is no longer just a luxury. It's infrastructure."
— Nicholas You, Executive Director, Guangzhou Institute for Urban Innovation, May 2026
What Smart Cities Actually Look Like Right Now
Let me give you a picture of where we actually are, because most coverage of smart cities oscillates between breathless hype and doom-and-gloom privacy panic. Neither is fully accurate.
Shanghai is probably the most ambitious example. The city runs a "Unified Management with One Network" platform that integrates transportation, emergency response, infrastructure monitoring, and public services into a single AI-driven system. The Shanghai Metro, one of the world's largest at over 800 kilometres and 400+ stations, uses AI monitoring, 5G communications, and predictive systems to reduce disruptions. Several lines already run fully autonomous, driverless trains. AI systems scan tunnels continuously for cracks, water leaks, and structural weaknesses. Residents can access over 3,500 public services through a single government portal.
Moscow operates one of the world's largest digital twin systems, updated twice a year via aerial photography, with over 9,000 analytical layers used to model traffic, infrastructure, and future development. Helsinki's Whim, a mobility-as-a-service platform, has led 38% of its active users to replace their daily car trips with multimodal transit options.
These are genuinely impressive. Real engineering, real scale, real results in specific metrics. And yet, despite all of this, something fundamental is missing in almost every smart city deployment I can find evidence of.
The UX Problem Nobody Is Naming
Here is the thing. When I talk to clients about designing enterprise SaaS products, one of the first questions I ask is: who is your actual user, and have you watched them try to use this? Ninety percent of the time, the answer to the second part is no. The product was designed around what the system can do, not around what the person needs to do.
Smart cities are doing exactly this at city scale.
The citizen is the user. The city is the product. And nobody is doing user research on that relationship. AI traffic systems optimize for throughput metrics that engineers care about. Smart surveillance systems optimize for detection accuracy that security teams care about. Digital twin platforms optimize for data richness that urban planners care about. What nobody is optimizing for is the actual experience of a person navigating the city on a Tuesday morning with two kids and a bus that is running late.
Gynna Millan, an architect and urban researcher from Colombia speaking at WUF13 in Baku, put it simply: technology in cities can either empower people or become "a tool for control," depending on how it is designed and who it is designed for. That framing is exactly right. And it is entirely a design question.
I have written about this dynamic in AI products before on Medium. The technology is rarely the failure point. The interface layer between the technology and the human is where things break down. Smart cities have the same problem, just at a scale of millions of users with no opt-out button.
Here is what a genuinely user-centered smart city design framework would actually include:
- Citizen co-design sessions before deployment, not after: The equivalent of user research sprints. Talk to the people who will live inside the system before you build it. This almost never happens. City governments and vendors design the system in closed rooms and then launch it. Residents find out about it when sensors appear on their street.
- Legible feedback loops: When an AI system affects you, you should be able to see how and why. A traffic rerouting decision. A transit fare adjustment. A permit denied by an automated system. Right now most smart city systems are black boxes to the citizens inside them. The UX design principle of transparency does not disappear just because the product is a city.
- Accessible interfaces across literacy and ability levels: Shanghai's portal offers 3,500 digital services, which sounds impressive. But if a 70-year-old resident with low digital literacy cannot navigate it, or if it requires a smartphone that a low-income resident cannot afford, those services do not exist for them. Accessibility is a first-class UX concern, not a compliance checkbox.
- Graceful degradation and override mechanisms: What happens when the AI traffic system makes a mistake? When the autonomous transit route gets it wrong? Every smart city system needs a human override pathway that citizens can actually access. Not a bureaucratic complaint form. An actual mechanism to flag errors and have them addressed in real time.
- Privacy by default, not privacy by regulation: Only 25% of smart cities currently conduct privacy impact assessments before deploying new technologies. That means 75% are building systems that collect citizen data without understanding the privacy implications first. This is not just an ethical failure. It is a product design failure. A product that erodes user trust is a broken product.
The Trust and Privacy Crisis Hiding in Plain Sight
Here is a number that should concern any product person working in this space: cyberattacks on smart city systems rose 30% in 2024. These are not just data breaches in the abstract. An attack on a city's AI traffic management system or its public transit control infrastructure is an attack on the daily lives of millions of people. This is exactly what Aytaj Khalafli, a cybersecurity representative from Azerbaijan, warned at WUF13: "If an attack happens, chaos can emerge in the country."
There is also a technical reliability problem that does not get enough attention. In Istanbul, centralizing AI inference for a smart city application added 80 milliseconds of network round-trip latency, which is enough to make a "real-time" system effectively useless for time-sensitive decisions. A system that works beautifully in a controlled demonstration and falls apart under real-world conditions is a UX failure. The user does not care about the architecture. They care whether the thing works when they need it to.
Who Gets Left Out When Design Is an Afterthought
The consequences of ignoring user-centered design in smart cities are not evenly distributed. They fall hardest on people who were already underserved.
Barcelona is a commonly cited smart city success story. The city has installed smart sensors across public spaces, AI-powered lighting systems, and connected infrastructure throughout the Superblocks project. All of that is real. But researchers studying Barcelona's smart city initiatives found that while the technology worked, it "did little to tackle deeper urban concerns such as housing affordability and wage inequality." The system was optimized for efficiency metrics. The actual quality of life for lower-income residents barely moved.
This is the defining failure pattern of smart cities right now: they improve the metrics that are easy to measure (traffic flow, energy consumption, sensor coverage) while leaving unchanged the outcomes that matter most to actual residents (housing, access to services, economic opportunity). Because the second category requires human-centered design. It requires understanding what people actually need from their city, not just what an AI system can optimize for.
Low-income and peripheral urban areas consistently show lower adoption of digital city services even when the physical infrastructure is available. The barrier is not always access to a device. Sometimes it is digital literacy. Sometimes it is trust. Sometimes it is language. These are UX problems. They require UX solutions: simpler interfaces, multilingual design, offline-first approaches, community-based onboarding. Cities are not investing in these.
The COVID-19 pandemic made this concrete in a painful way. In cities that had invested heavily in digital service delivery, residents of informal settlements were systematically cut off from education, healthcare services, and economic support because the entire service delivery system assumed digital access that those residents did not have. The smart city failed its most vulnerable users at exactly the moment they needed it most. That is a catastrophic UX failure.
What This Means for Anyone Building Digital Products
I spend most of my time designing enterprise SaaS products, AI-native interfaces, and agentic workflows. Smart cities are not directly my space. But everything I see in smart city design feels deeply familiar, because I have watched the same failure modes play out in product teams for years.
The pattern is always the same: engineering builds what it can build, without asking whether it is what users need. The technology gets deployed. Adoption is slower than expected. Satisfaction is lower than expected. Then someone calls it a "change management problem" or an "adoption problem" and tries to fix it with training or incentives. The actual problem was that nobody asked the users first.
Dmitry Maryasin of the UN Economic Commission for Europe said something at WUF13 that should be a principle for every product team: "Smart does not always mean introducing digital technologies or even innovation. Sometimes it is about good planning, setting priorities, and understanding how infrastructure is interconnected."
Replace "cities" with "products" and that sentence reads exactly like something I would put in a product strategy document.
The opportunity in smart cities right now is not better AI or better sensors. It is better design. The cities that figure out how to put citizen experience at the center of their technology decisions are going to see dramatically better outcomes, on every metric, not just the ones their engineers care about. And the firms and designers who figure out how to do that work are going to be very busy for the next twenty years.
What do you think? Are you seeing smart city technology in your city that works for you, or that fails you? Drop a comment below. I am genuinely curious what this looks like from the ground level in different parts of the world.
Sources:
1. UN News — Building the smart city: Promise, pitfalls and the people at its heart (May 2026) — https://news.un.org/en/story/2026/05/1167564
2. Smart Cities Council — Why Smart City AI Projects Fail: The Infrastructure Imperative — https://www.smartcitiescouncil.com/news/post/why-smart-city-ai-projects-fail-the-infrastructure-imperative
3. Phys.org — Why futuristic, tech-centered smart city projects are destined to fail (Feb 2026) — https://phys.org/news/2026-02-futuristic-tech-centered-smart-city.html
4. McKinsey — AI-native public infrastructure for smart cities — https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/tech-forward/how-ai-native-public-infrastructure-changes-how-cities-operate
5. StartUs Insights — 10 Emerging Smart City Trends 2026 — https://www.startus-insights.com/innovators-guide/emerging-smart-city-trends/
6. MDPI — Smart Cities, Digital Inequalities, and the Challenge of Inclusion — https://www.mdpi.com/2624-6511/7/6/130
7. GrayGroupIntl — Smart Cities in 2026: How Technology Is Reimagining Urban Life — https://www.graygroupintl.com/blog/smart-cities-2026/