Beyond the Hype: Building a Vision for Autonomous Vehicles (Part 1)
Key Takeaways
- To get the most out of autonomous vehicles (AVs), we need to develop a vision for how we want them to contribute to our communities.
- Too many of the visions for AVs are oversimplistic, saying that either AVs are good and so should become ubiquitous or bad and so should be heavily restricted. The truth is far more nuanced.
- To develop a fully rounded vision, we need to consider AVs from a wide range of perspectives.
- This blog looks at a subset of those perspectives (productivity, congestion, taxpayer funding and supporting technology development). Other perspectives will be considered in subsequent blogs.
- The productivity case for AVs is real but limited. While passengers can work during AV journeys, this doesn't offset the broader productivity benefits of healthy, active transport modes and efficient public transport.
- AVs could defuse congestion as a political issue. When people can work or relax while travelling, congestion matters less, creating space for long-overdue transport reforms like bus lanes and cycling infrastructure.
- AVs will require taxpayer subsidies without intervention. Road infrastructure, health costs from pollution and inactivity, and pressure for new pickup and drop off (PUDO) infrastructure all represent hidden subsidies that will burden taxpayers.
- The technology development argument doesn't justify regulatory inaction. Governments can support domestic AV industries with clear timelines for road user charging, providing both certainty and accountability.
What Next?
Subsequent blogs will examine additional perspectives for an AV vision.
Introduction
The conversation about autonomous vehicles has become frustratingly binary.
On one side, we have the techno-optimists promising utopia: ubiquitous shared AVs will end congestion, provide affordable transport for everyone, and make our streets safer, all whilst conveniently glossing over the unintended consequences in this brave new world.
On the other side, the response is equally simple: AVs are just cars, cars are bad, and it doesn't matter whether they're driven by humans or algorithms, petrol or electricity.
Both narratives miss the point.
Last year, I wrote about how we should regulate AVs to achieve better transport outcomes and made some rather pessimistic predictions about how we'd probably bungle their integration into our transport systems. I stand by those concerns. But rather than dwelling on what will likely go wrong, I want to spend this series exploring what should happen, what our vision for AVs ought to be.
This isn't an exercise in splitting the difference between the two polarised camps. I'm not interested in compromise for its own sake. Instead, I want to ask a more practical question: how can we leverage AVs to maximise benefits for our transport systems and communities?
I make no secret of my bias; I strongly favour people-oriented transport systems over car-dominated ones. But I'm also not going to pretend that AVs won't play a significant role in our transport future. They will. The question is whether we'll harness that role intelligently or squander it through regulatory inaction.
There are multiple dimensions to this challenge: how AVs function in dense urban cores versus suburbs and regional areas, their implications for freight and logistics, and their relationship with public transport. In this post, I'm focusing on dense urban areas and tackling the less glamorous but essential topics: productivity, congestion, taxpayer costs, and technology development. Future posts will examine accessibility, sustainability, fairness, and safety, as well as how the AV vision should differ across different urban contexts.
Let's start with productivity, a topic that advocates for people-oriented transport don't discuss nearly enough. Please note that I will do a separate blog on AVs and freight and logistics as the implications need a whole blog on their own.
Productivity
Productivity is the awkward topic that advocates for people-oriented transport systems tend to avoid, perhaps because it feels like conceding ground to the car lobby.
This is a mistake. Car advocates have long wielded productivity as their trump card, arguing that faster vehicle movement means more economic output. The utopian AV vision doubles down on this logic: machine-driven vehicles will eliminate accidents and smooth traffic flow, delivering a productivity bonanza.
But this argument rests on a fundamental misconception about how roads and productivity actually relate to each other.
Yes, if there's no road connecting point A to point B, building one can unlock significant productivity gains. But in dense urban areas, we passed that threshold decades ago. We're now deep into the territory of diminishing returns; adding more lanes and accommodating more cars produces ever-smaller productivity increases. Eventually, the curve bends downward. More lanes mean removing productive amenities (shops, housing, parks). More cars mean worse congestion. The result is a net loss in productivity.
AVs do offer one genuine productivity advantage over human-driven vehicles: passengers can work more productively during their journey. Answering emails or reading reports in an AV stuck in traffic is certainly more productive than gripping a steering wheel in the same jam. This could mitigate some of the productivity losses inherent in car-based transport.
However, this narrow focus overlooks that many people already work whilst on public transport, whilst walking and cycling confer cognitive benefits and long-term health improvements that enhance workforce productivity in ways that sitting in any vehicle, autonomous or not, cannot match.
And then there's induced demand. When you make driving easier and more pleasant, more people drive. I expect this principle to apply just as forcefully to AVs as it does to human-driven vehicles. The evidence for this will likely become undeniable over the next few years. In dense urban areas, this means AVs won't reduce congestion; they'll simply make it more tolerable for the people stuck in it.
The productivity calculus, then, is this: AVs represent an improvement over human-driven vehicles, and, in a very narrow sense, might even edge out properly designed public transport for work-while-travelling. But this ignores the broader productivity picture, the health costs of sedentary transport, the land consumed by vehicle infrastructure, and the time lost to induced demand.
Congestion
Congestion is politically potent because it's visceral. Drivers hate being stuck in traffic, the frustration, the wasted time, the sense of powerlessness. This emotional response drives enormous amounts of transport policy, usually in counterproductive directions.
But AVs fundamentally change the psychology of congestion.
If you're working through emails, watching a film, or reading a book in your AV, arriving 15 minutes late matters far less. You've already been productive for 30 minutes. The journey hasn't been wasted time; it's simply been mobile time. The frustration that makes congestion such a salient political issue could largely evaporate.
This psychological shift could be remarkably beneficial, but not for the obvious reasons. If congestion becomes less politically charged, it opens space for other overdue transport improvements. Politicians might finally feel emboldened to tackle stroads (those dangerous hybrid street-roads that serve neither function well), implement bus priority lanes, or reallocate road space to cycling infrastructure, changes that currently die in the face of "but what about congestion?" objections.
There's another potential upside: the ability to work whilst travelling might make people more willing to shift their journeys outside peak hours. If you're equally productive at 7:30 AM as at 8:30 AM, why not travel earlier and avoid the rush, especially if road-user charging makes it financially attractive? This could spread demand more evenly throughout the day, genuinely reducing peak-time congestion.
But these benefits assume we're proactive about regulation. Without intervention, the opposite could easily occur.
First, there's induced demand. Making vehicle travel more pleasant encourages more of it. If AVs make sitting in traffic bearable, more people will choose AVs over other transport modes. We could end up with the same congestion or worse, just with people who mind it less, hardly a win for our transport systems.
Second, and more insidiously, there's the pickup and drop-off (PUDO) problem. When everyone wants to be collected or deposited at the same locations during peak hours, you don't get smooth traffic flow; you get queues of AVs circling, waiting, blocking. We already have AVs stopping in bike lanes or bus stops. The result: pressure for yet more car-oriented infrastructure, vast expanses of kerb space and land dedicated to PUDO zones.
This would be tragic. One of the genuine benefits AVs offer is eliminating the need for parking infrastructure in city centres, freeing up enormous amounts of valuable land for housing, parks, shops, or simply better streets. If we replace parking infrastructure with equally sprawling PUDO infrastructure, we've squandered the opportunity entirely.
Our vision must be clear: AVs should reduce the political salience of congestion, enabling better transport policy overall. And they must not trigger demand for extensive PUDO infrastructure.
Get this right, and AVs become part of the solution. Get it wrong, and we've simply invented a new form of car-dependent city planning.
Taxpayer Funding
Transport funding is where ideology meets arithmetic, and the arithmetic is rarely what people expect.
For centre-right voters in particular, the question of "who pays?" matters enormously. Public transport's heavy subsidies are politically visible and frequently criticised. What's far less visible, and thus far less criticised, is that cars are also heavily subsidised by taxpayers, almost everywhere. People assume the various taxes they pay on vehicles, fuel, and registration cover the costs. Almost nowhere do they actually cover them.
So where do AVs fit into this picture?
At first glance, it's tempting to think AVs might be different. They're operated by private companies; surely that means less taxpayer subsidy? But this misses how transport subsidies actually work. AVs still use roads that must be built and maintained at public expense. They'll still need to park, at least some of the time. Someone pays for all of this, and it's usually taxpayers.
Then there are the indirect costs. AVs worsen health outcomes compared to walking and cycling. The resulting healthcare burden falls largely on taxpayers. While AVs being electric eliminates most tailpipe emissions, they still generate pollution from electricity generation, manufacturing processes, and tyre particles. Again, the health costs of this pollution are socialised.
But the bigger concern is what AVs might trigger in terms of government spending.
We're already seeing pressure to upgrade infrastructure to "support" AVs, smart traffic signals, dedicated lanes, etc. More worryingly, as I discussed in the congestion section, there will likely be intense demand for PUDO infrastructure as AVs multiply. Governments may feel politically compelled to accommodate this demand, building elaborate pickup and drop-off zones throughout city centres. All funded by taxpayers.
There's also the sprawl problem. If AVs make long commutes more tolerable, because you can watch Netflix or catch up on emails, people may accept living further from city centres. Urban sprawl is notorious for requiring taxpayer subsidies: the infrastructure costs of serving spread-out development rarely come close to being covered by the taxes those areas generate. More sprawl means more subsidy.
Now consider the policy environment you're trying to create. If you implement road user charging, as I believe we should, the number of private vehicles will drop, probably substantially. This will increase demand for public transport and generate significant revenue that can reduce public transport subsidies while improving service (last year, I wrote about what that might look like in New Zealand). Get the pricing right, and you can move toward a less subsidised transport system overall.
The conclusion for fiscal conservatives, or anyone who'd prefer taxpayer money go toward health, education, or welfare rather than car infrastructure, should be clear: you don't want AVs to simply replace human-driven vehicles. That path leads to continued or even increased subsidies.
Instead, the fiscally responsible approach is to ensure AVs are properly charged for their infrastructure use, to resist pressure for expensive PUDO infrastructure, and to prioritise public and active transport modes that can move more people more efficiently with less public subsidy, at least in inner-city areas.
The alternative is to sleepwalk into yet another generation of taxpayer-subsidised car dependency, just with fancier vehicles.
Supporting Technology Development
There's a geopolitical dimension to the AV discussion that governments find compelling: the desire to nurture homegrown AV technology companies. China, the United States, Europe, and the UK all share this ambition to varying degrees.
The logic is straightforward and seductive: create a supportive regulatory environment now, and your domestic AV companies will develop world-leading technology, compete globally, and generate jobs and prosperity at home for decades to come.
But this reasoning rests on several assumptions worth examining.
First assumption: short-term benefits will outweigh long-term costs.
This calculus only works if promoting AVs today doesn't lock you into poor transport outcomes tomorrow. If unchecked AV deployment leads to induced demand, sprawl, PUDO chaos, and continued taxpayer subsidies, all the problems we've discussed, then the jobs and tax revenue from AV companies may not offset the broader economic drag of dysfunctional transport systems. It's entirely possible to win the technology race whilst losing the livability and productivity contest.
Second assumption: AV technology will remain highly profitable in the long term.
This is far from certain. Over the next decade, dozens of companies may successfully develop autonomous driving. If they do, what happens to profit margins? AV technology could become commoditised, a competitive necessity rather than a source of sustained competitive advantage. In both the railway and airline sectors, the technology became available to many operators, and many companies became unprofitable.
So what should a government do if it wants to support domestic AV development and build better transport systems?
The answer is regulatory certainty with a built-in correction mechanism.
Create a welcoming environment for AVs now, streamlined testing approvals, reasonable liability frameworks, public-private partnerships for pilot programs. But simultaneously legislate that road user charging will apply to AVs in, say, two years. Existing road-user charging schemes, like those in London or New York, should apply from day one. But where schemes do not exist, give the industry a clear runway: two years of regulatory support to develop and scale, followed by pricing that ensures AVs contribute appropriately to infrastructure costs and don't simply generate induced demand.
This approach delivers several benefits. It provides regulatory certainty, which companies value enormously. It gives domestic AV providers time to establish themselves. But it also prevents the politically catastrophic scenario where millions of people start using unpriced AVs, making it virtually impossible to introduce charging later. Once people experience something as "free," charging for it becomes political suicide.
The timeline should be reviewed regularly; perhaps AVs need three years instead of two, but the principle is non-negotiable: don't wait until AV usage is widespread to introduce proper pricing. The window for establishing these rules closes quickly.
What about countries without significant domestic AV industries?
For these nations, the technology development argument evaporates entirely. There's no domestic industry to protect, no jobs windfall to justify regulatory forbearance. Some cities still tout their embrace of AVs as "forward-thinking" branding to attract talent and investment. But this is shallow thinking.
There are far better ways to position (and actually deliver) your city as innovative and livable: use AVs as a tool to improve your transport system and the quality of urban life, rather than as a branding exercise. A city with excellent public transport, safe cycling infrastructure, efficient road user charging, and well-managed AVs that complement rather than compete with sustainable modes, that's genuinely forward-thinking. A city that simply allows AV companies to operate with minimal regulation isn't innovative; it might provide a short-term βhighβ, but the come-down will be very messy.
The irony is that countries without domestic AV industries have the most freedom to get this right. Unburdened by pressure to protect local champions, they can focus purely on what serves their citizens and cities best. That's a competitive advantage, if they're willing to use it.
Conclusion
When we strip away the utopian fantasies and the reflexive opposition, a clearer picture emerges: autonomous vehicles represent neither salvation nor catastrophe for our cities, but rather a choice.
The analysis of productivity, congestion, taxpayer costs, and technology development all point to the same conclusion: AVs can deliver meaningful improvements over today's car-dominated transport systems, but only if we act now to shape how they're deployed. Left unregulated, they risk entrenching the same problems that plague our cities today: induced demand, sprawl, taxpayer subsidies, and the continued prioritisation of vehicles over people.
The window for getting this right is narrow. Once AVs flood our streets without proper regulation, the political cost of imposing road user charging or reallocating street space will become insurmountable. The time to establish the rules is now, before the technology becomes ubiquitous, not after.
In the next few posts, I'll explore how AVs should function in suburban, exurban, and regional contexts, examine their implications for accessibility, sustainability, fairness, and safety and consider their role in the movement of goods. Together, these pieces will build toward a comprehensive vision, one that ensures AVs serve our communities rather than the other way around.