The fully autonomous car—one that will carry you from point A to B with no human intervention—is coming. We’re not quite there yet; the National Highway Transportation Safety Administration (NHTSA) has laid out four different levels of autonomous car and even the most advanced adaptive cruise control systems on sale today represent only level 2 autonomy. But the technology for level 4 autonomous vehicles is not far off. Yesterday, Tesla Motors CEO Elon Musk told Ars that he thought Tesla would have the technology solved in about three years. Even the more cautious estimates we’ve heard from companies developing self-driving cars predict that they’ll be safer than human drivers within a decade. In fact, everyone in the industry that Ars has talked to agrees on one thing: the technology is going to be ready before we—society—are ready for it.

Before we start buying self-driving cars, people will need to be convinced that they’re safe and that they won’t be hacked or used to spy on us. Regulators will need new ways of satisfying themselves that the machines they’re allowing onto the roads are a safety improvement. And we’ll want to know who is liable for any collisions that happen when a car is driving itself. Last week we attended a debate hosted by Volvo and the Swedish Embassy in Washington, DC to delve into these topics.

Here in the US, safety is the main force pushing us toward self-driving cars. NHTSA Administrator Mark Rosekind gave a keynote address, telling the audience that 32,000 deaths a year on the road is unacceptable, particularly when 94 percent of them are attributed to driver error. “If technology will reduce deaths on American roads, [the NHTSA is] for it, right now,” he said. Rosekind also said that safety innovations should be pushed beyond the option lists of luxury cars and be available across the entire passenger fleet.

Read 4 remaining paragraphs | Comments