Just over an hour into Tuesday’s California Public Utilities Commission public meeting on the future of self-driving taxis, the machines took over.
“Please pardon the interruption,” a kindly robotic voice said, cutting into a government official’s prepared remarks. “Your conference contains less than three participants at this time. If you would like to continue, press star 1 now, or your conference will be terminated.”
In fact, there were three commissioners and two administrative judges sitting on the auditorium’s dais. The audience was full of industry representatives from Google sister company Waymo, General Motors, Uber, and Zoox, plus advocates from places like the Los Angeles Department of Transportation, the National Federation of the Blind, and the San Francisco Taxi Workers’ Alliance. They gathered to share their thoughts on the commission's proposed autonomous vehicle pilot program—the plan that would govern the way commercial self-driving services operate in the state.
Oh, and one commissioner had phoned in from Sacramento. She was, presumably, the robot’s concern.
For the Public Utilities Commission, the interjection exemplified the great fear: that technology does silly things, and when you load it with human cargo, silly quickly turns dangerous. The commercialization of self-driving tech is nigh. Waymo plans to launch a service in Arizona by the end of the year; General Motors’ Cruise will follow (somewhere) in 2019. The startup Drive.ai is set to launch a limited autonomous shuttle service in Texas in July. California began to approve special permits for totally driver-free vehicles in April. Which means it’s time for the Golden State to think about regulation.
That means running headlong into some of thorniest questions about driverless taxi services. What’s the safest way to roll them out? How should you regulate a technology that’s not “finished,” and never will be? What does the public need to know?
The commission’s answers, as presented by its proposed program, are not what companies like Waymo were hoping for. It wouldn’t allow companies to charge for rides, take passengers to or from the airport, or run shared, “pooled” rides. It would require any specific vehicle carrying passengers to first undergo 90 days of on-road testing. It would demand reams of data from developers: miles traveled, miles traveled without passengers (aka “deadheading” miles), collision and disengagement reports, and transcriptions of any communications between riders and driverless vehicles’ remote operators within 24 hours.
The pilot would last for an indefinite amount of time, to let the state government learn about what it might take to write final rules for self-driving taxis. But industry worries that if overly strict guidelines make it into this first stage, they'll eventually turn permanent.
If you’re wondering why the Public Utilities Commission is making these calls, it’s because it regulates transportation services like buses, limos, taxis, and ride-hailing, along with telecommunications and electric utilities. The Department of Motor Vehicles oversees autonomous vehicle licensing and testing. The federal government looks after vehicle design.
By dealing with the hard questions now, California is setting a template for other states, cities and towns, which will eventually have to grapple with this stuff. No wonder the industry is all hands on deck in California—where it will have a chance to set a bit of a precedent while charting the way forward in some of the country’s most valuable markets, like San Francisco and Los Angeles.
And they have to do it now: The commissioners are set to finalize and vote on their pilot project proposal next week. (It could be delayed if they decide to make significant changes.)
During Tuesday’s hearing, industry reps took particular umbrage with the state’s plan to force driverless taxis to first offer their services for free—how else are they supposed to figure out what consumers will pay?—and to nix pooled rides. “A prohibition on shared rides will prevent us from reducing single occupancy trips, traffic congestion, and greenhouse gas emissions,” said Laura Claster Bisesto, a senior public policy manager at Lyft.
The idea of collecting gobs of data (and possibly making it public) triggered similar consternation. “With the commission’s focus on consumer safety, that should be your filter when you’re asking, ‘What data do we need?’” Mark Rosekind, a road safety regulator in the Obama administration and now the chief safety innovation officer for the AV startup Zoox, said during Tuesday’s meetings.
Since a dude named Travis launched an app called UberCab in 2010, techy transportation companies have argued that providing regulators with too much data threatens rider privacy and their own intellectual property. “Data is the new electricity,” says Bryan Casey, a legal fellow at Stanford University's Center for Automotive Research. “The more tightly you can guard your data, the greater commercial advantages exist because of that asymmetry. You can put data out there at your convenience and public relations framing.”
Plus, that data can be misused or misinterpreted. California’s DMV, for example, solicits annual reports on disengagements, or moments when an autonomous vehicle’s safety driver must grab control from the tech. Critics say those numbers don’t illuminate much, and make those testing tech in tricky situations look bad. Industry reps also worry that handing over transcriptions of communications between AV’s remote operators and passengers could compromise private health information.
California’s regulators, for their part, are desperate to learn from the past. When Uber and Lyft rolled into the state a few years ago, the utilities commission had to create a whole new category of regulations for “transportation network companies.” By the time that rulemaking was under way, those companies had already swooped in and “disrupted” markets.
“The transportation network company experience has shaped everyone’s view of how to approach these things,” says Bryant Walker Smith, a lawyer who studies driverless vehicle regulations at the University of South Carolina School of Law. That was the classic horse-is-gone, close-the-barn-door maneuver, and this is an opportunity to start afresh. “If you’re going to open the barn door, make sure the horse is still in there and then open it a bit at a time.”
Regulators would like to ensure safety today while fostering an industry that promises even more road safety tomorrow—nearly 4,000 people die on California’s roads every year. But the balance is hard to find. Witness the plight of Arizona: The state’s governor welcomed autonomous vehicles with a no-strings-attached executive order in 2015. Following a fatal self-driving crash in March, he indefinitely suspended Uber’s right to test. This week, Uber shut down operations in the state, eliminating hundreds of jobs.
The industry does want to be regulated. Jumping through hoops give company’s tech credibility. “Saying you complied with an extensive set of regulations has never hurt anyone when things have gone badly down the road,” says Casey.
The history of technology, after all, is full of false starts. Back at the Public Utilities Commission hearing, the robot voice continued, reading out a long list of conference call menu options. "Well, this won't be powering any autonomy," someone said, and the room laughed.
- Sam Harris and the myth of perfectly rational thought
- How to send invisible messages with subtle font tweaks
- Can VR make porn profitable—and make the tech world respect the adult film industry?
- The inside story of the great silicon heist
- The Physics of a Tesla Model X towing a Boeing 787
- Looking for more? Sign up for our daily newsletter and never miss our latest and greatest stories
