By Sumit Paul-Choudhury Innovation is our regular column that highlights emerging technological ideas and where they may lead. Next time you ride an escalator, spare a thought for Bumper Harris – the one-legged chap supposedly employed by London Underground in 1911 to spend all day riding the newly installed escalators at Earl’s Court station to prove to nervous passengers that they were safe. Today we’re more likely to think of automation as cool than threatening, if we think of it at all. But public unease may return as machines programmed to follow instructions give way to “autonomous systems” that can learn, make decisions and take action by themselves. So the UK’s Royal Academy of Engineering argues in a report issued today, calling on the media and on government to improve public awareness of the complex social, ethical and legal questions that autonomous systems raise. But do we need this debate, or will it only delay the appearance of technologies that have a lot to offer? Two scenarios in the RAE report will be familiar to New Scientist readers: motor vehicles that drive, steer and park themselves, and “smart homes” that monitor health and boast friendly robots to provide companionship. Both offer substantial benefits. Autonomous transport could make roads safer, greener, less congested and more convenient. Health-aware homes and artificial companions could ensure that no one ever languishes without medical assistance or someone to talk to, something that could help richer nations cope with their rapidly ageing populations. But although these systems should be far more reliable than their human equivalents, people may not be willing to forgive their rare mistakes, said Will Stewart, visiting professor at University College London and the University of Southampton, UK, at the report’s launch. There may be a “yuk factor” when something goes wrong, he says, with the mistakes of an autonomous surgeon, for example, seen as inhuman and revolting in a way that human failings are not. Autonomous systems could also bring legal headaches. “The law is built around causes, and struggles with systems,” says Chris Elliott of Pitchill Consulting. The legal wrangling over which parties are ultimately responsible in today’s most complex cases, such as air crashes, could pale into insignificance compared with conflicts involving autonomous systems. If a driverless truck crashes into a human-operated car, who is to blame? The designer, programmer or manufacturer? Or even other vehicles that provided the truck with navigation information? Trying to pin the blame for a failure on individual parts or processes will be as meaningless as trying to decide whether it’s your right or left hand that causes the sound of clapping, Elliott argues. Those scenarios may be unsettling. But as Stewart said, “This is a new situation, not a newly dangerous situation.” Why whip up debate and potentially slow the pace of technological change when we have already assimilated many embryonic autonomous systems into our lives and accepted the consequences? Few of us wring our hands over the amount of thinking that cars already do on our behalf. Further incremental advances, provided they seem useful and safe, might be accepted without fuss, with the market steering innovation and legal cases mopping up when things go wrong. Making the public wary, on the other hand, might just result in pointless or counter-productive gestures. Figures like Bumper Harris could make a comeback. A laissez-faire approach may seem extreme. But the internet provides an example of how innovation that disrupts conventional thinking about responsibility, justice, morality and ethics can manage without paternalistic public debate. Can we afford autonomous systems the same freedom, or do they represent such a significant change they need special treatment? Read previous Innovation columns: Why do users fawn over Twitter’s failings?, Award-winning product design of 2009, Harnessing human nature to improve technology, When security meets surveillance, Physics brings realism to virtual reality, Smartphones need smarter networks, Looking forward to the smarter smartphone, How can Microsoft’s full-body gaming interface work?