POLICYMAKERS at every level of government are grappling with the question of whether fast-emerging autonomous-vehicle technology is safe enough for public roads.
Caution is a virtue, but there is also a cost to waiting. The sooner we get safe, usable driverless cars on the road, the sooner we can get drunk or distracted drivers off of the road.
About 94 percent of accidents are caused by human error. One in four involves alcohol.
We tend to be more comfortable with the devil we know, but humans are far more dangerous behind the wheel than the self-driving tech that already exists. Google's Waymo has driven more than 5 million miles in autonomous mode on public roads without a fatal accident.
Fortunately, self-driving cars finally have the green light for road testing in many states. Cities may be a different story.
California, for example, recently began accepting applications for statewide testing permits, yet San Francisco Mayor Mark Farrell requested that any company testing in his city participate in an additional program.
Given the recent tragedy in Tempe, Ariz., where an autonomous vehicle killed a jaywalking pedestrian, Farrell's request sounds sensible.
However, cities lack both the authority and the expertise to effectively regulate autonomous vehicle testing. By attempting to act as gatekeepers to their streets, they might actually limit the real-world testing that we clearly need.
First, while most cities lack the authority to stop companies from testing, the mere specter of an additional level of bureaucracy is sometimes a powerful deterrent. Similar issues arise when cities attempt to undermine everything from Uber to solar panels to wind turbines.
Previously, cities' autonomous-vehicle testing programs have been designed to make up for states' inaction. Pittsburgh, for example, touted its more streamlined process when Pennsylvania was slow to act.
But as states catch up and federal safety guidance continues to improve, fewer cities need their own processes. Going forward, local governments are more likely to create duplicative burdens that chase away innovators.
Second, it is unclear what, if any, value cities are adding to the process.
As Washington, D.C., transportation policy experts Ian Adams of the R Street Institute and Marc Scribner of the Competitive Enterprise Institute recently noted, "local jurisdictions simply are not qualified to evaluate what is 'safe' behavior when it comes to vehicle design, safety or performance."
Safety requirements for cars are regulated federally by the National Highway Traffic Safety Administration, with additional state regulations occasionally creating additional requirements.
San Francisco won't be the last major city to impose additional safety checks on autonomous vehicles. Unless states place clear limits on local governments, they'll wind up with a wild west-style patchwork of neighboring cities full of differing ideas about which types of cars are legal and which aren't.
If a large enough group of cities contradicts a state's laws, they may even prevent such innovation from developing anywhere in the state.
In that type of environment, many self-driving car companies may simply decide not to try. In response, Texas stated that local governments cannot impose additional regulations on autonomous vehicles.
Imagine how difficult driving would be if you didn't know if your driver's license was valid from city to city. Innovators face a similar dilemma when a city decides to undermine their "license" from the state with rules of its own.
It's worth noting that only one of the 50 companies developing self-driving cars in California, primarily in the San Francisco area, has applied for a permit so far.
No state wants to be left behind as others enjoy the economic and societal benefits of a technology that has the potential to save countless lives. It would be a shame for a few cities to spoil the role driverless cars can play in the future.