A policeman in Mountain View, California, pulled over a car recently for unusual reasons. The vehicle was moving at a measured pace, well below the speed limit and this aroused suspicions. When he got a closer look, the policeman was surprised to see there was no driving wheel or driver. It was a Google driverless car - not an uncommon sight in and around the headquarters of the information technology giant. Such automatic or semi-automatic cars could soon be commercially available with majors like Mercedes, Tesla, and BMW developing prototypes, apart from Google. A study by Business Insider Intelligence estimates that over 10 million driverless vehicles may be on the road by 2020. That will completely change the transport paradigm. While there could be many positive outcomes, there will also be difficult questions raised with respect to liability. Employment patterns will also change, up and down the value chain, as such vehicles become more popular. Google's cars are deliberately restricted to a sedate top speed of 25 miles per hour (about 40 km an hour). These prototype vehicles have, according to Google, logged around 1.9 million km, without a single traffic violation. What is more, every Google car receives the benefits accruing from the collective experience of that many kilometres on the road. The programming is tweaked to incorporate lessons as engineers working on the project learn from experience.
Automatic vehicles are undoubtedly superior to human drivers. The reaction times of driverless vehicles are much better than human reflexes. These cars obey traffic rules. They "see" via infrared and radar as well as normal light. Another very useful attribute is the ability to communicate with other automatic cars in real time. If a car knows when cars in front, behind, or approaching around a blind corner, are going to speed up, slow down, or change lanes, accidents may be drastically reduced. More vehicles can also drive with minimal braking distance at optimal speeds when all are aware of each other's actions. As wheels and gear boxes are removed, weight drops and space utilisation is optimised, leading to significant efficiency gains. Commuters will also enjoy enhanced productivity since they can work in comfort during their daily drives.
As of now, automobile accidents cause about 1.2 million deaths a year. The death toll could be cut to a fraction of that. Human errors will be reduced as drivers are eliminated and human errors are by far the most major cause of road accidents. But there will still be accidents. That is inevitable and some of them will be fatal, for sure. Accidents could be caused by glitches in software, by mechanical or equipment failure, or by some other factor such as errors by pedestrians or human drivers of other vehicles. Who picks up the tab for such accidents if there is no driver? Devising acceptable answers to that question could be the biggest stumbling block holding up widespread adoption of driverless cars. Insurers and vehicle licensing departments will need to think this through carefully before rewriting legislation to help assign liability. Apart from the car's owner, parties held liable could include the manufacturer or software developer (if that is a different party). Assuming that liability is reworked satisfactorily, as driverless cars and trucks catch on, there will be tectonic shifts in employment. Fewer professional drivers will be needed. Smart vehicles also "self-diagnose" problems to a much greater degree, reducing the need for traditional mechanics. But there will be enhanced requirements for frequent servicing by skilled technicians who review and upgrade the software, which must be protected from viruses and hackers. There will also be a market for redesigning vehicles to make them more productive and enjoyable.
It is clear enough that, one way or the other, driverless or semi-automatic vehicles will soon enter the market in large numbers. This will be truly disruptive. The benefits could be huge, in terms of safety, productivity and energy efficiency. But the regulatory system will have to cope with complicated questions of assigning liability to ensure that the potential benefits are fully realised.
