What are the policy challenges of driverless cars?

August 20, 2015

It was Jeremy Clarkson who offered one of the most memorable cautionary notes about the future of driverless cars. To paraphrase Amazon’s new TV star, “You have to remember that however sophisticated driverless cars get, somewhere there’s a guy called Keith tinkering under the bonnet thinking he can fix it himself.”

Whatever the risks presented by such hypothetical Keiths, the time of the driverless car seems to be drawing nearer. Advances in computing, sensors, big data analytics, GPS tracking – not to mention the cars themselves – have reached a point where the science fiction of the silver screen is inching closer to becoming the reality of the high street.

Policymakers need to think hard about what this might mean.

Two visions

Their starting point should be to recognise that there are two main competing visions for the future of driverless cars.

The first, proposed by the likes of Google (or should I say Alphabet?), is that the future is literally driverless. Humans – and the controls they rely on – will be taken out of the equation. In this model, cars need not look anything like those we know today. In place of the steering wheel and front-facing seats, passengers could have room to read, socialise, work, watch films or even sleep: a studio flat on wheels.

The second vision comes from automotive industry incumbents who believe the future lies in providing technological assistance to drivers. These cars will reduce the hassle and risk of driving: changing lanes, keeping a safe distance from vehicles in front, automatically braking to avoid danger, parking and navigating, but always requiring a person behind the wheel.

Which vision will win?

The answer is surely that there is room for both – but perhaps in different places.

Town and Country

The obvious application for fully driverless cars is in cities where they could become a major form of public transport. (For a taster of what this could look like, see below a clip of Heathrow Airport’s driverless pods that shuttle people between the long term car park and the terminals.) As pressure continues to grow on the UK’s urban areas, there is only so far traditional public transport (buses, tubes and trams, etc.) with fixed timetables and fixed stops can accurately match demand. To prove the point, Sir Peter Hendy, TfL’s former Commissioner of Transport, said “I predict that when Crossrail opens in 2018 it will be immediately full.” Driverless cars could be one answer.

Meanwhile, tech-assisted vehicles will surely be the model that thrives outside the city. Though people may hate the stop-start of commuting, there are many thousands who actually enjoy driving and are not yet ready to hand over all the controls. What else explains the prolific number of magazines, programmes and ads evangelising about the driving experience? Intuition suggests the majority of car owners are more likely to take the incremental step to a tech-assisted vehicle before making the giant leap to a completely hands-off version.

If these two worlds can coexist, what are the implications?

Start with the positives. Bringing fleets of driverless cars to cities could offer five big benefits:

1) They could increase labour productivity by moving workers efficiently to their exact destination. (How many hours are lost every day from people taking circuitous routes to work based on where the buses, tube trains or trams go?)

2) They could reduce congestion. Some studies (admittedly not uncontested) suggest that people looking for a parking space account for 30% of miles driven in US urban business districts.

3) As the UK battles to free up urban brownfield sites for housing, extra building space could be made available if fewer parking spaces were needed (enabled by replacing privately-owned cars that sit idle for 96% of the time, with shared cars that are in constant use). London alone has approximately 6.8 million parking spaces. If you subscribe to Ed Glaeser and Matt Ridley’s view that cities are the catalysts of innovation, increasing urban density is good for both jobs and prosperity.

4) Fixed stations inflate property prices around them (having a house near a Tube station in London was estimated in 2014 to add £42,000 to its value). Making public transport more ubiquitous with driverless cars could make it feasible for people to live in other areas, spreading the peaks and troughs of housing demand.

5) If they were battery powered, as per Heathrow, they could also help reduce air pollution, currently the cause of around 9,500 deaths in London each year.

Yet cities would face some big challenges, too.

First, they are likely to bear the brunt of job losses in some sectors. If there are concerns that companies like Uber are disrupting the work of established drivers now, we ain’t seen nothin yet. Travis Kalanick (Uber CEO) has publicly stated he wants Uber cars to be driverless in the future. As has been seen with discussions over automating trains on the London Underground, transport unions will surely object. But if prices are cheaper, expect the public to vote with their apps.

Second, cities will face the question of who should run the fleets of driverless cars. They could be owned / hired by city or local authorities, much as TfL owns Boris bikes. The alternative would be to let the private sector compete to offer their own services. Cities could encourage the best of both worlds if they set common standards in security, accessibility and ticketing, while letting the market dictate the prices. The latter may be preferable in insulating cities from the financial risk of finding which model works best.

Third – and whichever model is chosen – planners will need to get involved. Cities will have to find a way to work cooperatively with businesses to ensure the right infrastructure is in place to reap the benefits and mitigate the risks of driverless cars. This might involve offering driverless vehicles the use of bus lanes, or investing in new park and ride facilities where commuters using tech-assisted cars to reach the outskirts can transition to fully automated vehicles to get to their final destination.

There are also some big fears for the future.

Some have worried that thousands more jobs will be lost in haulage, when large freight becomes driverless. But the verdict is not yet so clear cut. It has long been possible for aeroplanes to be piloted by computers. That has not reduced demand for, nor de-skilled the role of, pilots. Passengers are simply not ready to place their trust entirely in a computer. View any lorry completing complex manoeuvres in high density urban areas, for example making deliveries near a school, and one suspects the same rules might apply – at least in the short term.

Both town and country will still have to address the big ethical question. If a driverless or tech-assisted car has to choose between killing its driver or ploughing into, say, a group of children, which should it decide? What are the implications for criminal responsibility, insurance, and so on?

For a while, this may be less problematic than it sounds. If tech-assisted vehicles remain in charge on the bulk of roads, the driver will surely still be responsible, just as they would be when using cruise control today. In the city – as with the Heathrow example – planners may choose to restrict the routes that can be taken by driverless cars, keeping them away from pedestrians, thereby limiting the risk of collision. Either way, as The Economist has pointed out, “regulators will demand that the algorithms that make such judgements should be audited, and their decision processes rendered transparent.” They will be right to do so.

As with all technological developments, only time will tell exactly what the impacts – both good and bad – will be. But one thing is clear. If UK policymakers want to harness driverless and tech-assisted cars to full advantage, they need to be thinking about the potential challenges and opportunities now.

Join our mailing list