Pratt in Tokyo
Who pays, or who possibly goes to jail when something happens is the biggest unsolved problem in the quest for the autonomous car, says a preeminent authority in the autonomous drive field. According to the expert, the infallible, 100 percent safe autonomous car is a fantasy. Autonomous drive will prevent a large number of accidents caused by inattentive humans, but we will have to come to grips with the fact that eventually, those robots driving our future cars will kill people – and then what?
Pratt and Toyoda
Today, Toyota announced a $1 billion investment into an R&D enterprise focused on artificial intelligence and robotics. Headquartered in Silicon Valley near Stanford University in Palo Alto, the company will be run by Dr. Gill Pratt, who led the DARPA Robotics Challenge when he was DARPA Program Manager. After today’s Tokyo press conference, I had a chance to sit down with Pratt, to discuss the future of autonomous drive, and to talk about what could stand in the way. The biggest problem is lawyers and the law, thinks Pratt:
“I think that liability is probably the number one issue in the autonomy field. When a car decides, the liability is with the car and the manufacturer. We don’t have the answers on how this will be done, but the question of who is responsible will make a huge difference here.”
Product liability has a huge impact on the auto industry. Automakers usually admit that indeed they have no easy answers to the liability question. The Silicon Valley camp usually glosses over the question, or it gives the pat answer that “insurance will cover it.” Many proclaim that with robots at the wheel, accidents will be a thing of the past anyway. Pratt vehemently disagreed today:
“These artificial intelligence techniques are very good, but the autonomy will not be perfect, it will not prevent all car accidents.”
Autonomous driving on highways with everybody going in the same direction, and no crossing traffic is quite doable, and should be commercially available soon, Pratt said. People who think that the car business will be disrupted by self-driving Uber pods, however, will require decades of patience. Pratt sees self-driving cars in inner cities only if there are special lanes, along with guarantees that there “are no rogue human driven cars to screw things up.” These lanes would basically turn cars into “a road train.” Once robots and humans mix on the roads, matters will get hairy. Says Pratt:
“The hard part is what to do with those crazy cars driven by human beings. If you do go to sleep in your autonomous car, and there is just a small chance that a human being will suddenly cut into the road, then that’s very hard to deal with. For this to become reality will take a much longer time.”
Around the world, some 1.2 million people die in traffic accidents each year, that’s over 3,000 people each day. They usually die at the hands of inattentive drivers, and society largely shrugs its shoulders. How will society react when the first person dies in an accident with a robocar? The answer will define the future of autonomous cars. Recklessly ignoring the question can stop autonomous cars before they become reality.
