A senator once asked the head of Google's self-driving vehicle program what sort of legislation was needed to help his industry. "What we have found in most places is that the best action is to take no action," he replied, adding that "in general the technology can be safely tested today on roads in many states."
Last week a congressional committee ignored that advice and took action.
In a bipartisan vote of 54–0, the House Energy and Commerce Committee has now forwarded the SELF DRIVE Act* for consideration by the full House of Representatives. (The Senate is working on similar legislation.) The bill's goal is to set up a national regulatory framework to encourage the development and deployment of autonomous passenger vehicles. But why does Congress need to get involved with autonomous vehicle development at all? After all, between 1900 and 1965 automakers managed to put tens of millions of non-self-driving vehicles on the road—some 90 million by the mid-1960s—with essentially no interference from the federal government.
The federal government didn't really get into the automobile regulation business until Congress created the National Highway Transportation Safety Administration (NHTSA) in 1966. The immediate impetus behind the push to create the new federal automobile safety agency was the publication of Ralph Nader's Unsafe At Any Speed: The Designed Dangers of the American Automobile, which claimed that GM's Corvair had a tendency to roll and therefore was a "one-car accident." In 1972, the very agency that Nader's alarmism conjured into existence issued a report finding that the "Corvair compares favorably with contemporary vehicles used in the tests…the handling and stability performance of the 1960–63 Corvair does not result in an abnormal potential for loss of control or rollover, and it is at least as good as the performance of some contemporary vehicles both foreign and domestic." In other words, federal automobile regulation was founded on activist misinformation.
Prior to 1966, automobiles somehow got fitted with such safety equipment as windshield wipers, headlights, and turn signals without federal intervention. (In 1939, Buick became the first U.S. automaker to offer factory-installed flashing turn signals.) Industry standards were generally devised not by bureaucrats but by the Society of Automotive Engineers.
On the postive side, the SELF DRIVE Act would preempt states from adopting their own rules for regulating "highly automated vehicles, automated driving systems, or components of automated driving systems." A year ago, the California Department of Motor Vehicles proposed a draft regulation that would require all self-driving cars to have steering wheels, pedals, and a licensed, specially trained driver in the front seat. Fortunately, the agency recently backed off those requirements. Nevertheless, 20 states have passed legislation related to autonomous vehicles and 33 states have introduced yet more such legislation this year.
But that's not all the bill would do. Among other things, it directs the secretary of transportation to issue within 24 months a final rule requiring autonomous vehicle manufacturers to submit safety assessment certifications; creates a Highly Automated Vehicle Advisory Council to undertake information-gathering activities, develop technical advice, and present best practices or recommendations to the Transportation Secretary; requires manufacturers to devise and submit cybersecurity plans; and prohibits manufacturers from selling highly automated vehicles until they have developed privacy plans with respect to the collection, use, sharing, and storage of information about vehicle owners and occupants. Perplexingly, the bill also protects state automobile dealer franchise laws that ban direct sales of cars. Since most autonomous vehicles will likely be operated as robotaxis, dealer franchises are likely to go the way of livery stables.
Curiously, the new bill requires that autonomous vehicles be as safe as conventional cars. Do folks in Congress really think that people would actually get in a robot car that they think is not at least as safe to ride in as a regular automobile? Just as conventional automakers affixed turn signals and windshield wipers to their vehicles, surely companies working on autonomous cars can be expected to add lidars, radar, cameras, GPS, and so forth without direction from regulators. Cybersecurity and privacy plans are great ideas—so great that autonomous car manufacturers are already addressing those issues.
This bill may be well-intentioned, but new federal regulations are more likely to hinder than to help the development of driverless vehicles. For more background, see my 2016 article, "Will Politicians Block Our Driverless Future?"
(* A tortured acronym: the Safely Ensuring Lives Future Deployment and Research In Vehicle Evolution Act.)