In a report released last week for CEI, I noted that developers need to be able to demonstrate automated vehicle safety benefits in order to justify releasing them to consumers. Based on road use and crash data, the critical milestone to reach is about 725,000 miles of crash-free driving in order to be 99 percent confident that automated vehicles are safer than manually driven ones.
Today, Chris Urmson, director of Google's Self-Driving Car Project, announced the company had reached nearly 700,000 miles of crash-free driving across its fleet of self-driving cars. Google's self-driving fleet has been involved in a couple of minor accidents, but they either occurred when the self-driving car was being manually driven or were due to the fault of another driver.
Google will be able to demonstrate safety benefits later this year when it hits the critical milestone. Once this occurs and research is published, expect even more private-sector investment and general public interest in automated vehicles to occur. But, as I've warned, this will also bring increased lawmaker and regulator attention.
As I detailed in my report, policy makers should take care to exercise great restraint in regulating automated vehicle technologies. Overregulation, particularly at this very early stage, risks locking in inferior first-generation technologies, delaying consumer availability, and increasing prices. The results of these good faith but foolish efforts would be increased property damage, injury, and death, as consumers are stuck in less safe, manually driven vehicles for longer than they otherwise would be.
The biggest dangers facing automated vehicles appear to be political rather than technical. There are many ways for lawmakers and regulators to screw this up. Let's hope they recognize their own limitations and allow what Mercatus Center Senior Research Fellow Adam Thierer calls permissionless innovation to occur.