With the unveiling of Google’s new car and renewed press interest in what autonomous vehicles might mean for society, much has been made of the unease many people have over unleashing computers into such a safety critical environment, where a computer controlled car might kill humans.
This ignores the fact that over 1M humans die on the roads each year across our planet, most caused by the failure of other humans to drive responsibly.
The future of autonomous vehicles is assured. Much of the required technology is already here. And production cars are already being fitted with the features that the cars of the future will need to drive autonomously.
We humans will have to suspend our fears about technology and once we accept autonomous vehicles on our roads, we will progressively have to aquiesce in further ethical conundra as the benefits of such vehicles become more obvious. Autonomous vehicles have the potential to free up our sclerotic transport systems in ways we can’t yet fully imagine.
Firstly, when taxis become autonomised, then we will become used to summoning an autonomous vehicle simply using a smartphone (much as we are already starting to do now) to the point where most of us will decide there is no point owning our own car. Why bother with all that maintenance and time spent trying to find parking spaces, when we become used to a car plucking us from place to place, efficiently, and navigating routes using predicted congestion data as it not only dodges current congestion, but also congestion that may be caused by other vehicles re-routing too. This will swiftly lead to far fewer vehicles being on our roads and less road space devoted to car parking, freeing up urban space for more beneficial uses.
Secondly, autonomous vehicles will be able to eke far more capacity out of current road space than we humans currently do, with our erratic and often less skillful driving habits. One can imagine convoys of vehicles travelling rapidly along a highway, each vehicle communicating with all others and with satellites, as well as detecting its environment – enabling the vehicles to travel at speeds of perhaps 160kph safely and with gaps of maybe only 2m between each vehicle. The vehicles will detect tidal flows of merging vehicles coming in from side roads and the stream will react, like a shoal of wheeling fish, with the new vehicles joining the convoy with barely an imperceptible slowing as they edge into their positions.
Consequences of this will be less car building and less road building – as we manage to live better more tranquil lives with fewer material inputs.
But ethically, future decisions for society will be contentious. Not far into the future, we will decide that human beings are just too dangerous and erratic to be trusted to drive vehicles at higher speeds on main inter-urban routes. And more controversially, our car-driving computers, currently programmed to avoid any collision or causing harm to any human, will be allowed to make decisions that will inevitably lead to people being killed. Consider our fast moving closely packed stream of vehicles travelling along a major highway – a human on foot deciding to try to cross that tide, for whatever reason, suicidal feelings or drunk or deranged, will cause huge danger to the occupants of the vehicles being whisked to home or holiday or work – and the computers will have to make decisions about which will cause least overall harm, recognising that some harm will have to occur – dilemmas that humans make poorly but which the computer-driven cars may perhaps make more rationally.
Autonomous cars promise much for society – making transportt far safer, cheaper and more convenient than it already is. But the step of deciding to trust cars to drive themselves is only a first and probably simplest one compared to dilemmas that may follow.
- Tim Chapman is infrastructure director at Arup