A robot may not injure a human being or, through inaction, allow a human being to come to harm.” That’s the first of Isaac Asimov’s three laws of robotics. But what does one do when a robot breaks a law? Reprogramme it, give it a timeout by switching it off for a while, or reduce it to scrap? In a hypothetical future where robots are animated by sentient, sapient AI, one can imagine the dread pooling in their metaphorical stomachs at the prospect of such a fate, knowing for sure that no afterlife awaits them — no wonder C-3PO in Star Wars is always so afraid of being destroyed. Perhaps justice will be reformative (reprogammative?) rather than retributive. And, in the case of an accident or mistake, does accountability lie with the robot, its owner — if there is one — or the manufacturer?
A recent incident in California offers a peek into the future. The San Bruno police pulled up a car for an illegal U-turn, but there was no one behind the wheel, and no passenger either. It was a driverless robotaxi from Waymo, one of the leading companies in autonomous driving. The Waymo Driver — the system that controls these vehicles — got off scot-free.
In this case, the police, after contacting Waymo, hinted that a reprogramming was in the offing and that hopefully, this would “keep it from making any more illegal moves”. But they might not be so lenient in the future — next year, a law will come into force in California allowing police to issue “notices of noncompliance” to the companies operating driverless cars for road violations. That closes the loophole for now, but there may come a time when the stakes are higher and the companies — or the robots — better at evading accountability.
