Autonomous Vehicles Are Coming—And We're Going to Mess It Up
Key Takeaways
- Autonomous Vehicles (AVs) will continue to roll out over the next decade.
- Here are 10 predictions for policy, infrastructure and regulation:
- Safety will cease to be an issue, but it will not be all smooth sailing.
- There will be a major cybersecurity incident.
- Most jurisdictions will NOT take proactive steps to manage how AVs integrate into the wider transport system.
- Public Transport will be negatively affected by AVs.
- Pick up and drop off will become a significant problem.
- Most AVs will not be shared.
- Autonomous Trucks will kill a lot of rail freight.
- Drivers will be protected in some places, but it will be temporary.
- Urban Sprawl Will Get Worse.
- Vehicle to Everything (V2X) Infrastructure will not be invested in.
What next?
Are you well prepared for AVs?
Introduction
Autonomous Vehicles (AVs) are coming. That's not a prediction; it's already happening.
The question now is what we'll do in response.
A couple of months ago, I wrote about how we could get the most out of AVs by implementing road-user charging and strictly adhering to a transport hierarchy. That article laid out what we ‘should’ do.
This blog is about what we'll ‘actually’ do.
Based on how most governments handle transport policy, I'm not optimistic. We've seen this movie before. In the 20th century, we redesigned entire cities around private cars without thinking through the consequences. We're now living with those mistakes.
Now we're about to do it again with AVs.
Most of the conversation around AVs focuses on safety and technology. Those are important, but they're not the whole story. The bigger questions are about how AVs will reshape our cities, our transport systems, and our lives. Will they reduce congestion or make it worse? Will they support public transport or kill it? Will they make our cities more or less livable?
This blog contains my top 10 predictions for AVs over the next decade. I'm not going to predict how many AVs there will be or which cities will adopt them first; there are people better placed to do that than I am. Instead, I'll stay in my lane and focus on what I know: transport systems, public policy and regulatory responses.
Spoiler alert: most jurisdictions will get it wrong.
Not because the technology is bad. Not because AVs can't improve our cities. But because we will lack the foresight to regulate them properly from the start. And once lots of people are using AVs every day, it will be too difficult to change course.
Prediction 1 - Safety will cease to be an issue, but it will not be all smooth sailing.
Let's get the safety debate out of the way first.
I'm not going to argue about whether AVs are safer than human drivers right now. AVs will continually get safer, unlike human drivers, and at some point will become demonstrably safer than humans, if they aren't already.
But AVs will never be perfectly safe.
They'll continue to generate high-profile, unusual crashes that humans would have easily avoided. And eventually, there will be a "trolley problem" incident.
An AV will face an unavoidable crash and have to make a split-second decision about who to put in danger. Maybe it swerves to avoid a child and hits an elderly person instead. Maybe it protects its passenger at the expense of a pedestrian. Whatever the choice, people will feel it was wrong.
The algorithm made a moral decision, and we'll all have opinions about whether it was the right one.
This will spark public debate. People will demand more regulation over AV decision-making.
So, we'll regulate.
Prediction 2 - There will be a major cybersecurity incident.
Almost any connected system has a risk of being hacked. AVs won't be an exception.
The nightmare scenario is a malicious actor finds a vulnerability in an AV fleet's system and takes control of thousands of vehicles simultaneously. They turn them into weapons, targeting pedestrians or critical infrastructure. It would be like 9/11, but with cars instead of planes.
Will this actually happen? I doubt it. I hope not.
The security around AV systems will be robust. Governments will mandate strict cybersecurity standards. Multiple layers of protection will be built in.
But perfect security doesn't exist.
What I do believe will happen is a smaller-scale incident. A hacker, will find a vulnerability and take control of a handful of vehicles.
Even a small-scale attack will have big repercussions.
The company affected will see its reputation severely damaged. People will be reluctant to use their vehicles.
The incident will slow AV adoption, but it won't stop it.
What will change is regulation.
Governments will respond with even stricter cybersecurity requirements.
But the rollout will continue.
Prediction 3 - Most jurisdictions will NOT take proactive steps to manage how AVs integrate into the wider transport system.
This is the big one. This is where we'll make our biggest mistakes.
AVs, if not properly managed, will have serious negative consequences for our transport systems. They'll increase congestion and they'll make our cities more car-dependent, not less.
We know this because we've seen it before with regular cars. We know what happens when you make car travel cheaper and more convenient.
The smart policy response would be to get ahead of this. Implement road user charging for AVs from day one. Prioritise public and active transport in policy and infrastructure decisions. Manage curb space carefully. Make the policy environment favour shared AVs over private ownership.
Some cities will do this. Probably the ones already committed to sustainable transport and people-oriented planning.
But most places won't.
Why won't they do more?
To begin with, the negative impacts won't be immediately obvious. Congestion won't spike overnight. By the time the problems are clear, lots of people will be depending on AVs for their transport needs and then it will be too late.
Restricting AVs will be politically toxic. Imagine trying to implement road user charging for AVs after people have gotten used to cheap, convenient AVs. Politicians will avoid it.
We'll have sleepwalked into another era of car dependence, except this time the cars won't have drivers.
There will be a handful of success stories. Cities that implemented strong policies early and stuck to them. They'll be held up as examples of what's possible.
But these will be the exceptions that prove the rule.
This is my most confident prediction. And the one that troubles me the most.
Prediction 4 - Public Transport will be negatively affected by AVs.
Ask the optimists, and they'll paint you a rosy picture of the AV future for public transport.
Autonomous buses will slash operating costs by eliminating driver wages, the biggest expense for many transport agencies. These savings will enable better service: more routes, higher frequencies, longer operating hours. Better service will attract more riders. More riders mean better cost recovery. Better cost recovery funds even more service improvements.
It's a virtuous cycle that leads to a public transport renaissance.
A few places will actually achieve this. Cities with strong political commitment to public transport and the institutional capacity to move quickly on new technology. They'll be the success stories everyone points to.
But for most cities, public transport will get worse, not better. Here’s why:
Firstly, robotaxis will directly compete with public transport, and they'll win. The convenience factor is overwhelming: door-to-door service, no waiting, no transfers, no crowds, climate control, privacy. For anyone who can afford it, and robotaxis will be very affordable, the choice will be obvious.
Public transport ridership will decline, potentially sharply.
Second, public transport agencies will be slow to adopt autonomous buses. Painfully slow.
Why? Because of the labour issue. Bus drivers are unionised, and unions will fight hard against automation that threatens their members' jobs. This isn't a criticism, it's their job to protect workers. But it will slow adoption significantly.
Even when agencies do start procuring autonomous buses, they'll move cautiously.
Third, when public transport agencies finally do adopt autonomous buses and start seeing cost savings, those savings won't be reinvested in better service, at best they will sustain existing low levels of cost recovery as patronage falls.
The result? A public transport system that serves fewer people.
Some cities will break this pattern. But most won't. By the time they realise what they've lost, it will be too late to get it back.
Prediction 5 - Pick up and drop off will become a significant problem
Picture this: It's 8:45 AM on a Monday in the city centre. Thousands of people need to arrive for work by 9:00 AM. Thousands of AVs are converging on the same few streets, all trying to drop off passengers at roughly the same time.
Now picture 5:30 PM. The same thing in reverse.
This is going to be a nightmare.
Curb space is already one of the most contested resources in city centres. AVs will make this problem much worse.
Cities will try to manage this. They'll experiment with different approaches: Designated AV pick-up and drop-off (PUDO) zones, queuing systems, etc.
These measures will help. Cities will get better at managing the flow. Technology will improve coordination between vehicles. We'll develop new traffic management systems specifically for AV PUDO.
But ultimately, there won't be enough curb space to handle peak demand efficiently.
So what will happen? Cities will build new infrastructure specifically to accommodate AV PUDO at scale.
In other words, we'll give even more urban space over to cars instead of people. We'll replace parking space with PUDO space. We'll trade one car infrastructure problem for another.
Some cities will resist this. They'll prioritise pedestrians, cyclists, and public transport in their curb management policies.
But most cities will build PUDO infrastructure to accommodate the demand.
We'll have solved the parking problem only to create the PUDO problem.
And we'll still be designing our cities around cars.
Prediction 6 - Most AVs will not be shared.
One of the most utopian visions of the AV revolution is that personal car ownership will be greatly reduced.
Why own a car that sits idle 95% of the time when you can summon a robotaxi whenever you need one? It'll be cheaper, more convenient, and better for the environment. Cities will need fewer parking spaces. Roads will be less congested because we'll have fewer vehicles overall.
It's a compelling vision, but it won't be realised.
Eventually (and possibly quite soon), someone, will successfully bring a private autonomous vehicle to market. And it will sell.
Why? Because people like owning cars.
Now, I know of at least one startup working on a model that could genuinely disrupt private ownership. If they execute well, they could capture significant market share.
But even in the best case scenario, they'll capture maybe 20% of the market. The majority will still choose private ownership.
We won't see the dramatic reduction in vehicle numbers. We won't free up vast amounts of urban space from parking, we'll still need most of it, just perhaps organised differently.
Some cities will buck this trend. Places with strong car-free movements and good alternatives. But they'll be outliers.
For most of the world, the AV era will look a lot like the regular car era: individuals owning cars that sit idle most of the time, taking up space.
Prediction 7 - Autonomous Trucks will kill a lot of rail freight
Rail freight has been losing ground to trucks for decades. Autonomous trucks will accelerate this decline.
Many governments have strategic objectives to shift more freight from road to rail. Yet despite these objectives, most countries have failed to shift freight to rail.
Why? Because road freight keeps getting more competitive. Autonomous trucks will be the final blow.
They'll transform the economics of road freight in two fundamental ways.
First, the obvious one: no driver means no driver wages. For long-haul trucking, driver wages are a significant percentage of operating costs. Eliminate that, and road freight becomes dramatically cheaper.
But the second advantage is even more significant: autonomous trucks can operate 24/7.
Human drivers have strict limits on driving hours, mandated rest periods, and practical limits on overnight driving. An autonomous truck has none of these constraints. It can drive straight through the night, when roads are empty and traffic is minimal.
This changes everything.
Delivery times will plummet. A route that takes two days with human drivers, accounting for rest stops and overnight breaks, becomes a single continuous journey. For many shipments, autonomous trucks will be faster than rail.
Overnight driving also means daytime roads will be less congested. Autonomous trucks will shift their operations to off-peak hours when roads are underutilised. This improves efficiency further and reduces one of trucking's traditional disadvantages: traffic delays.
Rail will be left with only the very longest hauls and the very highest volumes where the economics still work. Everything else will shift to autonomous trucks.
A few countries with strong rail traditions and government commitment might buck this trend. But most places won't.
Prediction 8 - Drivers will be protected in some places, but it will be temporary
Millions of people drive for a living. Taxi drivers, truck drivers, delivery drivers, bus drivers. When AVs arrive at scale, their jobs will be at risk.
How will governments and societies respond? Based on how automation has been handled historically, I predict three different approaches across different jurisdictions.
Approach 1: Do nothing meaningful
Most places will fall into this category. Drivers will lose their jobs and be expected to find work elsewhere with little or no specific support beyond standard unemployment benefits and generic job training programs.
Politicians will express sympathy. They'll announce modest retraining initiatives.
Then they'll move on to other issues.
Approach 2: Strong protections, but only for current drivers
A handful of jurisdictions, probably those with strong labour unions and progressive governments, will implement significant protections for existing drivers.
But these protections will apply only to current drivers. No new people will be allowed to enter these driving professions.
So you'll have a protected class of existing drivers who continue working, while the profession slowly shrinks through attrition.
This approach is politically palatable, it protects current workers without blocking technological progress indefinitely. It gives unions a win while allowing the transition to proceed on a timeline.
Approach 3: Genuine transition support
A very small number of jurisdictions will implement comprehensive transition programs that help drivers adapt to the changing economy.
Across all three approaches, one thing will be consistent: driving jobs will disappear to be replaced by new jobs related to new technologies and innovations.
Prediction 9 - Urban Sprawl Will Get Worse
There's a rule of thumb in urban planning that's held remarkably constant: people will tolerate about 30 minutes of commuting each way.
This "30-minute rule" has shaped cities for generations. It's why cities expanded with streetcars, then with metros, then with cars. Each new transport technology didn't reduce commute times, it expanded how far you could live while maintaining that 30-minute journey time.
AVs could completely upend this pattern.
The 30-minute tolerance isn't only about time. It's about the quality of that time.
Driving in traffic is stressful. But what if you're not driving?
In an autonomous vehicle, your commute time becomes your time. You can work remotely, taking calls and answering emails. You can read, watch shows, listen to podcasts, scroll social media.
Your car becomes a mobile living room, office, or entertainment centre. The commute transforms from wasted, stressful time into productive or leisure time.
When that happens, the 30-minute tolerance will expand. Maybe to 45 minutes. Maybe to an hour. Possibly more.
People will do the calculation: I can afford a much nicer house, if I'm willing to live further from work. And since I can work during my commute anyway, the extra distance doesn't really cost me anything.
All other things being equal, this will shift housing demand patterns.
Demand for properties in outer suburbs and exurbs will increase. Demand in inner-ring suburbs will soften.
The result will be increased pressure for urban sprawl and reduced pressure for densification.
This has serious consequences.
More sprawl means more car dependency, even with AVs. Outer suburbs can't support effective public transport, densities are too low, destinations too dispersed. Everyone will need a car (autonomous or not).
More sprawl also means more environmental impact as greenfield land is converted into suburbs.
Some cities will resist. But they'll be fighting against consumer preferences. It will require constant political will and enforcement.
Most cities won't have that will.
We'll look back in 20 years and see that AVs, despite all their promise, contributed to the same pattern we've seen since since the early 20th century: cities spreading outward, consuming land, increasing car dependency.
The technology will be slightly different. The outcome will be the same.
Prediction 10 - V2X Infrastructure will not be invested in
Vehicle-to-Everything (V2X) technology has been promised as critical for the future of AVs.
The concept is appealing: vehicles communicate with each other, with traffic signals, with road infrastructure, with pedestrian crossings. This communication enables smoother traffic flow, earlier hazard detection, coordinated intersection crossing, and improved safety.
However, it's not going to happen. At least not in most places.
Why? Because AVs work without it.
AVs can navigate complex urban environments using only onboard sensors.
Is V2X helpful? Yes. Would it make AVs more efficient and potentially safer? Probably. Is it necessary? No.
And "helpful but not necessary" is a terrible business case for expensive investment.
I'm already aware of research and startups developing solutions that significantly reduce or eliminate the need for V2X infrastructure.
As these alternatives mature, the value proposition for V2X infrastructure will shrink further.
Some places will invest in limited V2X infrastructure in specific use cases, such as high-value corridors (freight routes, transport priority lanes). But these will be boutique deployments, not the comprehensive infrastructure that V2X advocates envision.
Conclusion
If these predictions sound pessimistic, that's because they are.
But I'm not pessimistic about AVs. I'm pessimistic about our ability to manage them well.
The technology itself could be transformative. AVs could make our streets safer, our cities more accessible, and our transport systems more efficient. They could reduce the need for parking, free up valuable urban space, and give mobility to people who can't drive.
All of that is possible. But only if we regulate AVs properly from the start.
The problem is that "properly" means making decisions sooner rather than later, being proactive, not reactive. It means charging AVs for road use. It means prioritising walking, cycling and public transport.
Most governments make policies reactively, not proactively, so most of them will not make these choices.
Instead, they'll take the path of least resistance. They'll focus on safety regulations, which are important but insufficient. They'll allow AVs to roll out without thinking through the wider impacts. And by the time the problems need to be addressed, it will be too late to change course.
A handful of cities will get it right. They'll be the ones already committed to people-oriented transport planning. Those cities will show us what's possible.
But they'll be the exceptions, not the rule.
For everywhere else, AVs will be another chapter in the same old story: promising technology, poor planning, and cities designed around cars instead of people.
We could do better. We should do better. But I don't think we will.
What do you think? Am I too pessimistic? Or are we really about to make the same mistakes all over again?