When Nico Larco went for a morning run through the streets of Prague before the conference, he was struck by their typical medieval intertwining. In his view, Prague should not change this very urban landscape because of autonomous cars. Larco knows what he’s talking about; he heads the Urbanism Next center at the University of Oregon and has dedicated numerous studies to the research of autonomous mobility. Geographer Federico Cugurullo, who lectures at the National University of Singapore and Trinity College in Dublin, has also been studying how driverless cars are changing cities for a long time; for example, he and his team investigated the deployment of self-driving cars in Wuhan, China.

Autonomous vehicles have been with us for about ten years now. What lessons have we learned from the past decade, and what mistakes have we made?

NL: There was truly great enthusiasm, euphoria, and optimism surrounding autonomous vehicles, but these expectations were not met. We are far from achieving the level of adoption we anticipated. Part of it is a technological problem, but part of it is an economic challenge—self-driving cars are expensive. They work really well in densely populated areas where the starting and destination points are close together, i.e., in city centers. However, it turns out that these areas usually offer many other transportation options, so perhaps that’s precisely where the new technology isn’t needed. Conversely, the areas where it would make the most sense are located further from the center, but it is difficult to find a sustainable business model for them. For this to work, much more capital and vehicles are needed.

How has the rollout progressed in Asia, where you work, Federico?

FC: Over the past three years, my team and I have conducted a series of field research projects in Wuhan to see what we can learn from this city’s example. The difference from the American context is that in China, the government oversees large-scale artificial intelligence projects, and autonomous vehicles are part of them. At the same time, there are also municipal governments that set the conditions for their deployment and actively adapt the infrastructure to accommodate them. In cities like Wuhan, the infrastructure is being adapted for autonomous vehicles by integrating sensors into it. This way, cars perceive their surroundings not only through their own cameras and LiDAR systems, but also through sensors embedded in the city itself. Imagine it this way: the vehicle doesn’t just have its own “eyes”, but also uses those around it. In a sense, it thus becomes part of the city as a broader interconnected artificial intelligence system. However, this is only possible because China does not take a liberal approach to urban development.

There are two approaches, but both are far ahead of Europe. What should Europe learn from the American and Chinese models?

NL: My advice would be to test extensively to determine what this technology is capable of. People in the private and public sectors and in research are still just guessing what its large-scale deployment will look like. So I would recommend that Europe focus more on pilot projects. However, I would say that one of the great strengths is that Europe has essentially said: “Technology must help us with social and environmental problems and the issue of equality. Companies can do whatever they want, but if they can’t handle these things, we’re not interested.” In other words: test the possibilities, but then make sure the technology is actually deployed in a way that helps you achieve the desired outcome.

FC: If you want to deploy autonomous vehicles on a large scale, you need planners at the helm who collaborate with private companies. You can’t leave planning entirely to the commercial sector, because it will logically prioritize its own profit over the public interest.

What do you think are the most common misconceptions about driverless cars?

NL: One of the biggest, yet often overlooked, issues is cost. The deployment of autonomous vehicles depends on it to a large extent. This technology isn’t cheap, even though prices are currently falling—for example, LiDAR prices have dropped significantly. And then there’s the issue of insurance and liability, which could ultimately prove even more expensive. No one can say today how much this technology will cost in ten or thirty years. But if autonomous cars became cheap enough to be more affordable than today’s cars, that would be a fundamental change.

FC: Another misconception is that there will be massive resistance to self-driving cars. I don’t believe that will be the case. I think that once people try them, most will be fascinated at first and then find it boring.

NL: Another common misconception is about how people will use their time, which I’ve been researching with my team. It’s often assumed that autonomous vehicles will become a space for leisure. You’ll get in and spend time there with friends or family. You’ll even be able to have a drink if you want, or catch some sleep. In reality, however, studies show that in the 21st century, most of us are close to burnout and are essentially working all the time. And that’s how it will be in an autonomous vehicle, too. It won’t be a time of inactivity; you’ll be working in it.

How should cities like Prague, with their network of narrow streets, sidewalks, and various modes of transport, prepare for autonomous vehicles?

NL: I went for a run through Prague this morning; this is my second time here. It’s a beautiful city, and its beauty is partly due to that medieval street grid, those narrow alleys. The worst thing you could do is destroy these features for the sake of cars. Prague should prioritize walking, cycling, and public transportation. And then there could be specific areas where self-driving cars operate. For example, it’s important to designate zones for boarding and alighting; otherwise, it will be very chaotic.

FC: Urban planners in general shouldn’t get too caught up in the hype surrounding artificial intelligence, because it may not be beneficial when it comes to Prague’s uniqueness. This applies generally. If you read the scientific literature carefully, you won’t find much evidence that artificial intelligence on its own makes cities more sustainable. After all, you don’t want to base your policy on assumptions and hypotheses.