The death of self-driving vehicles has been greatly exaggerated. That was the message of two experts, on the subjects of artificial intelligence and transportation, who recently weighed in on the matter.
“I think the pendulum of public opinion has now swung too far in the pessimistic direction,” said Timothy Lee, author of the newsletter Understanding AI. “Self-driving technology has steadily improved over the last few years, and there’s every reason to expect that progress to continue.”
SENATE AIMS TO NAVIGATE CONFLICT BETWEEN COPYRIGHT AND TRAINING AI
Lee pointed out that companies Waymo, created by Google, and startup Cruise, principally owned by General Motors, “have continued to plug away at the problem.” These two firms presumably “don’t believe self-driving technology is ‘decades away’ because they’re already testing it in Phoenix and San Francisco,” Lee wrote. He was countering recent news reports that would suggest otherwise.
Money is one problem. After many confident predictions in the last decade that such vehicles would be the wave of the future, most of the investment that has gone into self-driving, or autonomous, vehicles has not seen good returns. Uber and Lyft both abandoned their AV projects. Ford and Volkswagen closed a joint venture.
Another unanticipated problem may be congestion, which AVs had been predicted to make better. A study released in May by the University of Copenhagen’s Department of Computer Science determined that the few driverless cars currently on the road could be making congestion slightly worse, not better, because computers don’t pick up on the social cues that drivers do to determine when to yield.
Public opinion is another hurdle. A survey released in March by AAA found that things are seemingly only getting worse. “This year, there was a major increase in drivers who are afraid, rising to 68% as compared to 55% in 2022,” the news release accompanying the survey said. “This is a 13% jump from last year’s survey and the biggest increase since 2020.”
There is also added press scrutiny when things go wrong with AVs. For instance, a Waymo taxi hit and killed a dog in San Francisco in May. “We send our sincere condolences to the dog’s owner,” a Waymo spokesperson told the Guardian, one of the many news outlets around the world that covered an occurrence that is commonplace, though awful, among the world’s 1.4 billion licensed drivers of nonautonomous vehicles.
Federal and local red tape can be another problem. For instance, Fairfax County, Virginia, recently conducted a two-year experiment with a self-driving “Relay” electric shuttle in the Mosaic shopping district. It performed as hoped for over 350 trips but was recently taken off the route.
Virginia officials say the AV “Relay” shuttle was a success but have not committed to going wider with the program. Why not? Transportation analyst Marc Scribner sketched out the bureaucratic hurdles to self-driving transit in a recent issue of the Reason Foundation’s Surface Transportation newsletter.
“Section 13(c) of the Urban Mass Transportation Act requires that local agencies accepting federal transit grants,” the vast majority of them, “must implement protections for existing employees,” Scribner explained.
True, the UMT Act “allows the elimination of jobs, but only as workers presently holding those jobs retire or vacate the positions for other reasons,” a 1976 report from the Office of Technology Assessment said. “Thus, the economic benefits of workforce reduction through automation of an existing transit system may be deferred for a number of years until retraining, transfer, or attrition can account for the displaced workers. Alternatively, direct compensation can be paid to affected workers, eliminating the jobs earlier but at an earlier cost.”
Scribner gave an example of how this law has greatly slowed down the automation of America’s transit systems and could make transit AVs a harder sell. “Automation of subway systems has been possible since the early 1960s,” he said, and yet most transit agencies haven’t been willing to pay those short-term costs for longer-term savings.
Still, Scribner does see some promising AV projects on the horizon.
“In April, the San Jose City Council voted to authorize a predevelopment agreement (PDA) for a proposed 3.5-mile airport connector project between Mineta International Airport and the Diridon Station rail hub in downtown San Jose,” he reported about the Silicon Valley city in northern California. “What makes this project unique is its planned use of automated shuttles developed by Glydways, which would operate on elevated guideways with a footprint width of a bicycle lane. The pairing of novel automated vehicle (AV) technology from Glydways with established infrastructure developer Plenary Americas suggests a path forward for public transportation applications of AVs.”
The savings could be substantial at a time when transit agencies are trying to figure out how to cut costs once federal stimulus dollars dry up.
CLICK HERE TO READ MORE FROM THE WASHINGTON EXAMINER
It’s also possible that this project would not run afoul of the federal statute, given the unique nature of these self-driving transit vehicles, or Glydcars. The vehicles hold only four sit-down passengers, so the argument could be made that they are not, individually, displacing bus or other transit drivers. Self-driving Glydcars are also being pitched as easy on rider wallets.
“One of Glydways’ principal selling points is that it claims the system can be built and operated profitably by charging passengers no more than prevailing transit fares,” Scribner said.