Waymo’s self-published safety benchmarks continue to best those recorded by human-operated vehicles, but a new spate of events is showing its imperfections.
San Francisco is the epicenter — and, by proxy, a testing ground — for autonomous driving. Self-driving test cars and robotaxis, some of which look like glorified air fryers, weave through the seven-by-seven, racking up tens of thousands of real-world driven miles and booking figures that rival (or even surpass) monopolized ride share companies.
But there are clearly some kinks in the proverbial iron. It’s become increasingly clear that robotaxis are not immune to pranks and human-like oversights.
Earlier this month, some fifty autonomous Jaguar I-Paces were left hobbled after each was individually summoned to San Francisco’s longest dead-end street.
Videos and images posted on X on October 12th show dozens of the Waymo-owned vehicles efficiently listless, confused about how to act next. Per X user Riley Wataz, it all began with a single vehicle called well before sunrise … and then another … and then another, each summoned by someone. It was a real-life, group-effort DDOS (distributed denial-of-service) attack, a tactic commonly seen in cyberattacks; because of Waymo’s no-show policies, each ping resulted in a $5 no-show, but the cars still stuck around for at least 10 minutes before driving away.
“They left after about 10 minutes and charged a $5 no-show fee,” Wataz wrote on X, later noting that the Google-owned company “handled [the issue] well” — “I assume this isn’t much different than if a big concert had just ended,” Wataz continued.
However, the group’s DDOS meant rides were disabled in 2 2-block vicinity until later in the morning; no traffic laws were broken by the group, as well.
Earlier in October, a parked and unloaded Waymo was involved in a small fender bender when a Lexus UX seemingly left a garage … and immediately struck the car. The Waymo vehicle was temporarily parked in front of the garage, with about half of it jutting out of the garage entrance and exit ramp. However, while the self-driving vehicle was clearly obstructing the area, it’s clear the driver of the human-operated Lexus failed to yield; it appears Waymo was stationary when it was struck; a picture of the incident uploaded to Threads shows an on-duty SFPD officer speaking to at least two individuals who presumably were involved in the small fender bender.
“Human vs robot in San Francisco,” reads the post uploaded to the Meta-owned platform. “Dammit, Wyamo,” the post continues. Sure, the driverless car could’ve moved out of the way — one could argue if it was operated by a human, it would’ve done so — but it’s the driven vehicle’s failure to yield that appears to be the cause of the accident.
The most severe, dangerous, human-like incident was in late September when an operating Waymo sped right past a stopped bus, filled with kids.
It’s been reported that a Waymo-operated robotaxi passed a stopped school bus in Atlanta that was unloading children at a neighborhood stop. The bus, which was stationary and had deployed its flashing stop sign, was passed by a Waymo robotaxi when it decided not to heed the stop and, instead, drive around it. Per the San Francisco Chronicle, the National Highway Traffic Safety Administration (NHTSA) has opened up a case and is gathering records for review; NHTSA has yet to comment if it will prosecute Waymo, though past probes into the company have closed with little ramifications.
Artificial intelligence is booming, all while simultaneously hitting an industry-wide collapse. Waymo’s success is intrinsically tied to A.I.’s future; its faults and cracks and flaws, too, are just as connected. We humans are responsible for the technologies that continue shaping the world and, ironically enough, must remove our own blind spots in those products to ensure a safer world — be it behind or in front of the driverless wheel.
