When a woman was hit and killed by a self-driving Uber taxi in Arizona recently, it was both an awful shock and a horrifying inevitability.
The crash was the combination of a series of tragic and small problems. If the car had been going slightly more slowly, if either its human or computer driver had spotted the woman slightly earlier, or if the road was slightly differently designed, then it’s possible no accident would have happened.
But there was also a grim lack of surprise in the crash. The cars were always destined to be involved in fatal accidents – even if they are projected to happen much less than they do in cars driven by humans – and that will continue, with this first death all but guaranteed to be followed by many more.
Limiting those fatal crashes will be the responsibility of the companies, policymakers, designers, engineers and other experts. But it could also be the responsibility of a group little talked about in the debate over autonomous vehicles: town planners, and the politicians who decide how we build the places we live.
There was always going to be a first crash. But this one seemed in so many ways to exemplify the challenges those concerned by the social impact of autonomous vehicles now face: the woman hit by the car was pushing along her bike, a form of transport mostly neglected by many modern city planners, and some have suggested she was forced to cross in the middle of the road by a road scheme that is almost actively hostile to pedestrians.
That the car was part of a self-driving vehicle trial by Uber, a company known for providing cabs, might appear strange. But the race to make self-driving cars is being entered by just about everyone with any connection to automobiles: traditional car companies, technology firms such as Google and taxi companies like Uber. (That the latter two industries have always had something of an allergic reaction to regulation might be chilling to the people who have to share roads with their vehicles.)
Those technology companies have done something like this before. Autonomous vehicles are often compared to the internet: both potentially transformative, both potentially disastrous. And getting them right could require avoiding some of the mistakes that were made when the web was introduced.
“What we’re seeing is that the focus has been largely on the engineering side – on developing vehicles,” says urban planner Mark Wilson. “What we found with other technologies like the internet, and will probably find with autonomous vehicles, is that the social context lags way behind.”
Those social questions are not about engineering. They’re about how we share and live in our cities: how space should be taken by cars and road, and how those cars should be controlled. It’s not too grand to say this might require entirely rethinking what we expect from a city.
“On the dream side, we’d see a road with vehicle-sharing, more efficient vehicles, perhaps a denser city and more efficient use of urban land area,” says Wilson. “Then there’s also the planner’s nightmare: people own autonomous vehicles and won’t share, they drive more miles, people will live in the countryside and sip coffee and work on their laptop while the vehicle travels even further, because now the time and frustration has been reduced.
“We all bring our own hopes and dreams. For planners, we are optimistic but it still may not come about.
“The internet started as a very optimistic technology, but what we’ve seen is that it’s bogged down in great complexities related to how it’s used.”
If we’re not careful, self-driving cars might simply make it easier to get around; they’ll trigger no deep rethinking of our cities, apart perhaps from letting people travel further and longer and thereby entrenching many of the problems that currently flow from unequal and inefficient transport systems. “When they invented low-fat foods, people ended up eating more of them,” says Jarrett Walker, a public transit consultant who regularly speaks about the future of transport.
“The problem, the challenge of induced demand in particular is enormous.”
Now is precisely the time those questions need to be asked: someone has died on our streets, and did so in service of a technology that remains mostly hypothetical to most people. “Not many people have shared a road with an autonomous vehicle, or at least known about it,” says Wilson. “We don’t have a lot of insight into how people will feel – whether it’ll be upsetting, off-putting or reassuring. We just don’t know yet.”
But if most people believe autonomous vehicles to be a long way off, many of the people who should be thinking about it appear to think those questions have already been answered.
“Very frequently, when I’m talking about any issue about how to build a city or to build a better transport system in the pre-automation condition, many people have been trained to tune out and not care, because soon we’ll have autonomous cars and they’ll change everything,” says Walker.
“There are two problems. First, the cars won’t change everything: there are still basic geometric problems. And secondly they undermine support for a space efficient, large-scale public transport system.”
On his website, Walker describes how he came to do the work he does: seeing the transformation of Portland, Oregon, as the city changed direction and adopted public transit. “Experiencing this transformation as a teenager – commuting by bus across the city through a downtown that grew more vibrant by the day – taught me to believe in the possibility of rapid and fundamental change in how a city imagines and builds itself,” he writes. It’s easy to imagine teenagers in a few years will be living through the exact same kinds of revolution, for better or worse.
Whether we integrate autonomous vehicles into a public transit network, or an even more extreme version of the current, privatised network of road vehicles, is another question that governments must yet still have mostly not reckoned with. Uber’s work on self-driving vehicles is largely selfish – it hopes eventually to do away with the drivers that it has to pay, and who have largely served to be a frustration because they ask for proper rights and protections.
But Uber’s car will at the very least be shared; if it is successful, it might do away with some of the paradigm of each person owning a car, which is unsustainable as cities continue to grow and become more dense. “The only way that we can make an autonomous vehicle work is if they are fleets – if you experience them as taxis rather than cars in your driveway,” says Walker.
This is only halfway towards the utopia that some thinkers lay out, in which those autonomous vehicles will operate more like buses.
“We have to compare the autonomous car with autonomous public transport,” says Walker. “There simply will be no way for public transport to compete with their sudden abundance and the reduction in the hassle of driving.
“On the other hand, the sudden increase will cause so many problems that it will be an emergency.
“I assume that when we have abundant and scalable driverless vehicles, we will also have driverless buses. And because public transport operating costs are mostly labour, it can be massively easier and cheaper to run.
“That’s the only way the autonomous future works.”
The world has addressed something like these problems before. Cities had run for thousands of years before the car turned up, but even the most ancient places were able to adapt to the new vehicles. It’s easy to forget, however, just how much work that took and how much time was spent; there’s much less labour and time available this time around.
“The first automobile required constant innovation,” says Wilson. “Someone had to invent the stop sign, or the traffic light.
“We also had an an incredible willingness in many places to embrace the new technology and change our way of life in order to accommodate it – neighbourhoods were bulldozed, motorways were built.”
There appears to be very little appetite for restructuring our cities today. Even the places that have welcomed autonomous vehicle firms into their cities – Arizona, where the Uber crash happened, has welcomed the trials with open arms, which self-driving carmakers like in part because there is no complicated weather to get in the way of sensors – haven’t made any great moves towards fixing transportation. The Uber crash happened on a road that was built for cars, and a world controlled by them, and autonomous vehicles are mostly being placed on those same roads with few adjustments.
“In an ideal world we would overlay our planning interests with transportation. Transportation has always been a force in designing cities,” says Wilson, pointing to the fact that cities were once designed only for walking, then for public transport, and now for cars – though in ways that still need to map onto those old walking paths.
“Many cities, especially in the UK, are hundreds of years old or even thousands of years old and they’re not designed for traffic. There are problems with retrofitting cities into modern footprints.”
We may finally have a chance to reverse some of that trend. We might not need quite so much road space, for instance – so what if that became cycle space, or footpaths? “Each one of these decisions is going to be subject to national and local forces that are trying to influence the outcome, and one of the challenges in all of this is who has the responsibility for educating the public, meeting the public and explaining to voters what the future holds,” says Wilson.
“While the engineering and automotive world has been looking at this for quite a while, it’s fairly new in terms of politics, and planners and the social context.”
Join our new commenting forum
Join thought-provoking conversations, follow other Independent readers and see their replies