It will soon be easy for self-driving cars to hide in plain sight. We shouldn’t let them.

0

It will soon become easy for self-driving cars to hide in plain sight. The rooftop lidar sensors currently marking many of them are likely to get smaller. Mercedes vehicles equipped with the new partially automated Drive Pilot system, which carries its lidar sensors behind the front grille of the car, are already indistinguishable by the naked eye from ordinary human-driven vehicles.

Is this a good thing? As part of our Driverless future project at University College London, my colleagues and I recently concluded the largest and most comprehensive citizen attitude survey autonomous vehicles and the highway code. One of the questions we decided to ask, after conducting more than 50 in-depth interviews with experts, was whether self-driving cars should be labelled. The consensus of our sample of 4,800 UK citizens is clear: 87% agree with the statement “It should be clear to other road users if a vehicle is driving itself” (only 4% were disagree, others unsure).

We sent the same survey to a small group of experts. They are less convinced: 44% agree and 28% disagree that the status of a vehicle should be announced. The question is not simple. There are valid arguments on both sides.

We could argue that, as a matter of principle, humans should know when they are interacting with robots. This was the argument made in 2017, in a report commissioned by the UK Engineering and Physical Sciences Research Council. “Robots are manufactured artifacts,” he said. “They must not be deceptively designed to exploit vulnerable users; instead, their machine nature should be transparent. If self-driving cars on public roads are genuinely tested, then other road users could be considered subjects in that experiment and would have to give something like informed consent. Another argument in favor of labeling, this one practical, is that – as with a car driven by a learner driver – it is safer to give wide berth to a vehicle which may not behave like one driven by a well-trained human.

There are also arguments against labelling. A label could be seen as an abdication of responsibility for innovators, implying that others should recognize and welcome an autonomous vehicle. And one could argue that a new label, without a clear shared sense of the limits of technology, would only add confusion to roads already full of distractions.

From a scientific point of view, labels also affect data collection. If a self-driving car learns to drive and others know and behave differently, it could alter the data it collects. Something like this seemed to be on the mind of a Volvo executive who told a reporter in 2016 that “just to be on the safe side”, the company is reportedly using unmarked cars for its self-driving test project on UK roads. “I’m pretty sure people will challenge them if they’re marked by really hard braking in front of a self-driving car or getting in the way,” he said.

Overall, the case for labelling, at least in the short term, is more compelling. This debate is not limited to self-driving cars. It gets to the heart of the question of how new technologies should be regulated. Developers of emerging technologies, who often represent them as disruptive and changeable at first, are likely to describe them as merely incremental and trouble-free once the regulators knock on the door. But new technologies don’t just fit into the world as it is. They reshape worlds. If we want to realize their benefits and make good decisions about their risks, we need to be honest about them.

To better understand and manage the deployment of self-driving cars, we need to dispel the myth that computers will drive like humans, only better. Management professor Ajay Agrawal, for example, argued that self-driving cars essentially do what drivers do, but more efficiently: “Humans receive data through the sensors – the cameras on our face and the microphones on the sides of our heads – and the data comes in, we process the data with our monkey brains and then we take action and our actions are very limited: we can turn left, we can turn right, we can brake, we can speed up.

Share.

About Author

Comments are closed.