It’s all or nothing for self-driving cars
image by Victor Gonzalez Couso on Flickr

Companies like Tesla, Google, and Uber suggest that autonomous vehicles may hit the road in the next five years. Self-driving cars may save us a lot of time, but it remains questionable whether authorities and the general public will embrace these vehicles. Will we ever be able to celebrate that ‘only’ ten thousand people died in a self-driving car?


  • The first, remote-controlled, driverless vehicle drove in Ohio in 1921. It was followed by the ‘American Wonder’ that drove on Broadway in 1925.
  • The GM-sponsored Futurama exhibit, at the 1939 N.Y. World Fair, featured families in self-driving cars on an ‘Automated Highway System’. In the 1950s GM actually built concept vehicles, e.g. the Firebird II, that could follow electromagnetic wires in the road surface.Each year, over 35,000 Americans die in traffic. 94% percent of all accidents are due to human error and at least 10% of fatalities can be attributed to distracted drivers. After decades of steady decline, traffic fatalities have risen again this year, most probably due to drivers engaging with their smartphones.
  • Tesla will include all the hardware necessary for autonomous driving in its upcoming Model 3. So far, however, its Autopilot function still requires the human driver to pay attention and intervene when necessary.
  • According to Google, the only way to make AVs safe is to eliminate the role of human drivers altogether. Trials with semi-autonomous cars showed that test drivers eventually stopped paying attention and were unreliable partners for the vehicles.


Developers of autonomous vehicles (AVs) disagree on the road to commercialization of their cars. Google has pledged that it will only develop, and market, fully autonomous cars that relieve human drivers of all duties. Most others, including Audi and Tesla, opt for a more gradual approach by aiming first for semi-autonomous vehicles in which humans are still expected to pay attention and intervene when necessary.

Fully autonomous cars will take longer to develop, but will bring about clear and valuable benefits and have a good chance of increasing traffic safety

The dilemma is clear. Semi-autonomous vehicles can be commercialized much sooner than their fully autonomous counterparts, but their added value is limited and they may very well prove less safe than today’s cars. By contrast, fully autonomous cars will take longer to develop, but will bring about clear and valuable benefits and have a good chance of increasing traffic safety.

While the current debate on AVs focuses mostly on saving time and on more effective use of existing road capacity, there is perhaps too little attention for their potential to save lives. By contrast, the earliest ‘self-driving cars’ were used to underscore the importance of driving safely. That is, these driverless cars abided by the rules and caused no incidents and if humans did the same, they would neither.

Given that the AV is promoted primarily as a means to save drivers’ time or as a nice-to-have, we are unlikely ever to accept fatal accidents like the kind Tesla experienced earlier this year. Drawing inspiration from the first driverless cars, the focus, perhaps, ought to be on eliminating human error and saving lives. One can imagine that car makers prefer not to talk about fatalities at all, but if AVs do successfully conquer the market we may be able, in twenty years’ time, to celebrate the happy fact that ‘only’ ten thousand Americans were killed in traffic while driving their autonomous car.