Driving Me Crazy

As if driverless cars weren't exciting enough, we are now told they could be a target for hackers intent on causing chaos. This was spun by some media outlets as the promise of terrorism, which shows that the government line that uncontrolled Internet access is a threat to society has been successfully internalised. According to the UK-based Institution of Engineering and Technology, "If we have the hacker community start to target vehicles, we can imagine a fair amount of chaos". Why only a "fair amount"? Let your imagination run wild. I'm thinking remote-controlled carjacking, like a mashup of Grand Theft Auto and Police, Camera Action!.

Most of the press coverage has focused on the equation of hacking with disruption (and not in the business-bollix sense of that word): "Unfortunately living in the world today people do try to tamper with technology". The phrase "living in the world today" is grammatically redundant but it helps suggest that there is a happier, purer realm where this sort of bad behaviour does not occur. What this patronising twaddle ignores is that most technological progress is the result of hacking and tampering, i.e. experimentation and repurposing. For example, the Internet is the result of various academics repurposing spare computer time and network bandwidth.

It is significant that the car-hacker meme should prove popular in Britain. To be hackable, a driverless car must support two-way communication with a surrounding infrastructure - i.e. it has to be capable of taking instruction. As the name suggests, a fully-autonomous car (which is what many American advocates want) would not be controllable in this way, in which case hacking would require physical intervention (the digital equivalent of hotwiring). The IET, which is a pro-infrastructure lobby, reckon: "There are clear safety benefits in getting vehicles to drive optimally and at a particular speed while interacting with the infrastructure. We're moving towards co-operative systems. The first trial will happen in the UK in January. Vehicles and traffic lights will be talking to each other – that's the first step." Presumably the second step will be for GCHQ to listen in on the conversation.

There was also ample evidence of the IET's dubious understanding of technology in its comments on the software that driverless cars will use: "Recent reports analysing software show that 98% of applications have serious defects and in many cases there were 10-15 defects per application ... If ultimately you want to use autonomous vehicles, we need to make sure they don't have a defect". That headline figure relates to all software, not to critical packages whose failure might imperil life. It's also worth pointing out that "defect" is a broad term that covers more than just bugs. It includes trivial feature deficiencies and cosmetic preferences. In terms of hacking, an exploitable vulnerability (which is what you need to worry about) could be the result of a bug, but it's more likely to be the result of bad design (e.g. an open backdoor) that isn't classed as a defect. The IET are advocating state regulation of the software. This is not a bad approach per se, but it raises questions about which state agencies will be involved.

Some media picked up on the research finding reported by the IET that in a mixed environment manual car drivers ended up copying driverless cars, leaving shorter braking distances and thus causing more accidents. Again, this has a particular ideological resonance in respect of the relationship of the individual and the infrastructure: we must protect you from yourself. This backs up my own theory that driverless cars will only happen as a result of government diktat and strict segregation, because the chief benefits, such as increased road capacity and fewer accidents, can only arrive when automated cars are mandatory and the non-automated outlawed within a particular area, which almost certainly means high value city centres.

Transport policy is rich in hyperbole because of its emblematic associations with progress (railways, jet packs) and personal autonomy (the freedom of the road). In the US this encourages preemptive "libertarian" attacks on federal regulation ("say no to Obamacar"). In the UK we get the man from the ministry annex warning about cyber-carjacking and trying to persuade us that what we need is a virtual chauffeur under the control of the state. To get a better idea of what the reality of driverless cars might look like, it's worth employing three traditional memes that the British media regularly recycle, which tell us a lot about national attitudes to transport policy.

1. The US hates pedestrians --- According to Ben Okri, "There are cities in America where you can’t get anywhere if you don’t have a car". This myth is the flipside of the British habit of regarding pavements as a human right, to the point of treating dogshit as a political totem. Of course, it didn't stop us building plenty of new towns and exurban estates without pavements over the last 50 years, just like the US did. I spent my teenage years in Washington, near Sunderland, where desire paths crisscrossed the "highway" verges (pavements? what luxury!). In fact, the two countries (and developed nations generally) are very similar in terms of urban transport policy, which has been increasingly pro-pedestrian since the emergence of New Urbanism in the 90s.

Walking is a class issue. During the postwar era in the US, getting about on foot came to be seen as the recourse of the poor. I recall walking in car-friendly (but getaboutable) Houston in the late 80s and noting that my fellow pedestrians were mainly Latino maids knocking off from a shift at the hotel I was staying in. In the UK, we cultivated that attitude to an extent (the middle-class signifier of the driving coat and open-back gloves, like car-keys at a swingers' party, was big in the 70s), but still retained a residual respect for walking as a democratic practice, which probably reflected the UK's greater and longer-established urbanism and the less extensive "white flight" of the 60s and 70s. The "US hates pedestrians" meme was a self-congratulatory boast on our part.

As the suburbs have lost their sheen and inner-cities have been gentrified, walking has been reclaimed by the middle classes (the recuperation of psychogeography and the dérive in the 1980s was an intellectual outrider of this tendency). Together with the vogue for cycling, this has significantly increased transport costs because of the need to provide multiple routes and segregation in built-up areas. The response has been a trend towards shared spaces, which is why driverless cars are often pictured as non-threatening golf-buggies, pootling along as they carefully avoid cyclists and pedestrians.

2. Traffic speed in central London is no faster today than in the Victorian era --- This curio is often taken as evidence that we're technologically stagnating and might as well still be using horse-drawn carriages. In fact, it has long been known that the average traffic speed in urban areas is determined by the minimum that people will tolerate, not by technology, and that this speed is 9mph (roughly three times normal walking speed and double "hurrying" speed). This is self-regulating. When congestion increases and the speed drops lower, people avoid travelling and congestion eases. If congestion decreases to the point where the average speed exceeds 9mph, this attracts more traffic and congestion increases.

Even if we had jet packs, we'd still be travelling in town at 9mph (which means we'll need to wait for anti-gravity suits as jet propulsion is shit at low speeds). Given that driverless cars can most easily be introduced in city centres, this suggests that vehicles designed to be optimal at around 9mph will be the order of the day. They will be able to go faster, but I suspect they'll max out at 20mph as a reassurance to pedestrians and cyclists. It is also doubtful that car capacity will increase significantly, because of mixed use constraints, so city centres will remain as they are today, areas that are dominated by the cars of the rich.

3. We're a small island --- This is a flexible meme that is deployed in a variety of arguments, from limiting housebuilding to reducing immigration. We are in fact the ninth largest island on Earth. Most of the larger ones are covered in ice (Greenland, Baffin Island), impenetrable mountain-ranges (New Guinea, Borneo), or highlands unsuitable to dense habitation (Madagascar). There are essentially two large "sweet spot" islands that combine rich agricultural soil, easily-accessible resources and proximity to trade routes: Great Britain and Honshu, Japan's main island. The most habitable bit of the former is England, suggesting that Cecil Rhodes' remark about life's lottery has a material basis even if it is an objectionable sentiment.

The unthinking acceptance of this meme is largely a result of city-dwellers (and particularly Londoners) feeling cramped and the residents of small towns in England not getting out enough. A theoretical selling point for driverless cars is that they can allow for an increase in effective road capacity (though I doubt this will happen in practice). In other words, the acceptance of autonomous vehicles can be presented as an alternative to additional road building or widening. Though this is largely spurious, it is an effective way of selling driverless cars that simultaneously encourages the belief that they are particularly suited to congested urban areas.

What I think the collective lens of these memes shows is that driverless cars will initially be little more than a futuristic version of the sedan chair. Found largely in city centres, employed mainly by the well-off, and possessing the properties of a limousine: quiet, smooth, and with a built-in screen and minibar. The malicious hacking threat is not that a terrorist might try and turn you into a suicide ram-raider, crashing into the Downing Street gates at 20mph, but that you might be inundated with video-spam.