The Drum Awards for Marketing - Extended Deadline

-d -h -min -sec

Highway Code Ford USwitch

Slow progress: Can the public be convinced that driverless cars are safe?

Author

By Cameron Clarke, Editor

March 19, 2015 | 6 min read

Much like 19th century fears around the first motor vehicles, 43 per cent of people say they don’t trust driverless cars. Cameron Clarke finds out that there’s still some way to go in convincing the public to cede control to machines.

The first cars struck so much terror into Britons that parliament passed a law decreeing that they could travel through towns no faster than two miles per hour – slower than the speed at which the average person walks. The Locomotive Act of 1865 further stipulated that these alarming new machines had to be operated by three people: two to drive, and one to walk ahead waving a red flag as a warning.

That we are now contemplating a future of cars with no drivers at all is a testament to the relentless progress of technology and our acceptance of its place within our lives. But for all our modern enlightenment, some of those 19th century fears about ceding control to machines are being rekindled.

According to research from uSwitch published this month, 48 per cent of consumers would be unwilling to be a passenger in an autonomous vehicle, 43 percent wouldn’t trust a car to drive safely without a driver and 16 per cent are ‘horrified’ by the idea. As the driverless car prototypes begin their trials on UK roads, it is obvious that while the technology is getting closer, there is still some way to go to convince the public they are ready for it.

Peter McOwan, a professor of computer science at the Queen Mary University of London, is sympathetic. “For me, anything that starts to change the way we live our lives for the better is fantastic… [but] as someone walking down the side of the road, I can share some of the concerns.

“I think what’s frightening people is it’s suddenly a de-skilling. On the whole, people who drive enjoy driving, and people who drive think they’re good drivers. So if somebody’s going to come along and take that away from you, there’s a combination both of ‘you’re taking away something that I enjoy and I don’t like that’ and ‘I don’t know how safe it is’.”

The Department for Transport (DfT) report ‘Pathway to Driverless Cars’ lists safety as a primary benefit of driverless cars. “Human error is a factor in over 90 per cent of collisions,” the report reads. “Failing to look properly, misjudging other road users’ movements, being distracted, careless or in too much of a hurry are the most common causes of collisions. Automated vehicles will not make these mistakes.”

If that makes them sound infallible, McOwan, whose research interests include robotics and artificial intelligence (AI), knows that the way systems work in a laboratory is no guarantee of how they will behave when it matters.

“When you move them out into the real world there can be issues like rain, sunlight, fog – and what happens if one of the beacons at the side of the road fails? You need to see how they work in real environments with real people getting in, getting you would never really have thought of doing in the lab.” And as McOwan points out, like any connected technology, driverless cars will have to overcome another vulnerability: hackers.

What the DfT’s report overlooks is that behind every driverless car, there are human beings responsible for programming it to behave the way it does. When collisions are unavoidable, drivers have to make value judgements. Do you stay in the road and hit the pram that’s veered in front of you, or do you swerve onto the pavement and plough into the middle-aged couple walking arm in arm? How will a computer deal with that?

Scott Le Vine, a transportation planner and research associate at Imperial College London, believes such moral dilemmas will place a huge burden on those responsible for programming vehicles’ AI. “On the one hand it is desirable from a societal point of view if we have a system that has fewer deaths,” says Le Vine. “But the family and estate of someone who does get killed are going to look for someone to sue and they’re going to analyse either the explicit ethical coding in the algorithm or the lack of it. If I were designing those algorithms, I would see that as a tremendous risk.”

In simplistic terms, the thinking behind driverless cars being safer is that because they would follow the Highway Code implicitly, and not fall prey to human failings like texting at the wheel, they will cause fewer accidents. But there’s a real likelihood that travelling in such an obedient and, let’s be honest, slow machine might be regarded a monotonous experience.

“There’s an awful lot of ambiguity in traffic laws and there’s an awful lot of ambiguity in how human drivers deal with the guidance we’re given,” says Le Vine. “We exceed the speed limit; these vehicles in principle should not be disobeying the speed limit. If I’m Ford or General Motors, and I’m thinking about programming every one of my thousands and tens of thousands of vehicles to break the speed limit, as soon as there’s an accident there’s going to be a terrific lawsuit.

“I don’t think we’re going to allow that to happen – that’s silly. I think what will happen is we’ll end up adjusting traffic laws. Moses didn’t come down from Mount Sinai with speed limits on a tablet saying ‘thou shall drive at 70 on a motorway’. If you read the speed limits as they are today they’re rigid, but they’re flexibly interpreted. I think they’ll become more flexible in how they’re written.”

Which is, of course, exactly what happened to the Locomotive Act. By the time it was withdrawn in 1896, the speed limit had been increased to 14mph. In time, it became even more relaxed as the law adapted to the technology.

But we’re a long way off needing to worry about that, according to Le Vine. “In terms of putting your 12-year-old in the car to go off to football practice by themselves, I wouldn’t envision that happening in the 2020s. A car that drives itself everywhere, every time, under every set of circumstances... that, clearly, we’re decades away from. What we’ll see is these technologies creep up, rather than a driverless car appearing in the showroom one day and you having to decide to buy it or not.”

Just like those drivers of the late 1800s, we’ll have to make do with slow progress for now.

This piece first appeared in the 18 March issue of The Drum. To buy a copy or to start subscribing go to The Drum Store.

Highway Code Ford USwitch

More from Highway Code

View all

Trending

Industry insights

View all
Add your own content +