Weekly Brief: Autonomous Tech Under Safety Microscope
The California Public Utilities Commission may suspend or revoke Cruise’s driverless permit in San Francisco following accusations that the company has overlooked safety issues in its race to commercialize a robo-taxi service.
The accusations originated in an anonymous letter from a whistleblower who claims to be an employee at Cruise who works on safety critical systems. The whistleblower asserts that Cruise’s self-driving technology has a pattern of getting confused at intersections and bringing traffic to a standstill until Cruise employees show up on site and manually tow vehicles out of trouble. This has happened with such regularity, says the whistleblower, that these episodes have earned the internal moniker of VREs, or Vehicle Retrieval Events. “Employees generally do not believe we are ready to launch to the public but there is fear of admitting this because of expectations from leadership and investors,” wrote the whistleblower in his letter.
Cruise responded by touting its safety and reporting records, noting that 94% of its employees agree that safety is a top priority at the company, per an internal survey conducted in April. This doesn’t look good for anyone. The CPUC received the whistleblower letter back in May, before it issued Cruise’s final driverless permit, but didn’t agree to an investigation until after Cruise launched its robo-taxi service and shut down a city block in San Francisco for several hours within its first week of service. I reported on that incident here. The CPUC still didn’t open an investigation until the Wall Street Journal acquired a copy of the letter and let the commission know that it would be taking the story public. Not a good look for a government agency.
Cruise has been zealous (perhaps overzealous) in its drive to beat Waymo onto the streets of San Francisco and claim the prize as the first commercial robo-taxi service in a major US city. The positive press of claiming that prize may now be overshadowed by the negative press of hiding safety issues in order to achieve it. Uber learned that lesson the hard way in 2018 when one of its self-driving vehicles struck and killed a pedestrian in Tempe, Arizona. Its whole self-driving vehicle program fell apart thereafter, offering a cautionary tale for how quickly the pendulum can swing when public perception and support turns against you.
That may not happen here. The whistleblower may be grossly exaggerating reality; Cruise may never lose its permit, making this a blip on the start-up’s glide path toward nationwide expansion. Or it could be the start of something far more damning. Despite the fact that Waymo has more experience and more tested technology than any of its competitors, it has opted to take a slower, less headline-splashy route toward commercialization. Hence the reason that it first commercialized in Phoenix, Arizona, and why it has taken its time in San Francisco. If Cruise stumbles, Waymo is squarely in the driver seat of America’s robo-taxi future.
In other news last week, Tesla’s senior director of artificial intelligence, Andrej Karpathy, is stepping down after five years at the company. Karpathy is respected in the AI field as a visionary but Autopilot has struggled to take the next step toward full autonomy and has come under increasing scrutiny from the US federal government for its involvement in a string of fatal accidents. Two weeks ago Tesla laid off nearly 200 Autopilot employees. The departure of Karpathy raises more questions about the future course for Tesla’s self-driving aspirations.
