Editor's Note: San Francisco officials are currently permitting an experiment with self-driving cars, an incredible technology in America’s capital of innovation, and a miraculous leap in mobility that should be proudly celebrated. Of course, it’s San Francisco, so instead we’re locked in a destructive, hysterical, activist war for…? It’s never entirely clear, but in this case it does look like labor union power.

As ever, California is just our nation’s prologue. The introduction of working, autonomous vehicles to American streets is about to catalyze a ferocious, partisan war. Sanjana Friedman and Brandon Gorrell break the story down for Pirate Wires.

-Solana

---

At the end of 2021, California began to issue permits for SF-based autonomous vehicle (AV) companies Cruise and Waymo, which are both now beta testing rideshare services. The rollout, we’re told, has been an unmitigated disaster. In June, San Francisco community leaders sounded the alarm after one of Waymo’s vehicles killed a small dog. A week later, the Guardian reported one of Cruise’s vehicles obstructed emergency crews responding to a mass shooting. Last week, a Twitter user posted a video of his “near death experience at the hand of a driverless car” after a Waymo “ran a yellow/red,” and he was “nearly jackknifed.” Now, lobbied by parent companies Google and GM, state regulators are poised to lift all AV restrictions, granting the companies unfettered, 24/7 access to our streets, and the ability to start collecting fares. In a last-ditch effort to save their city from robotaxi hell, SF city officials, agencies, and concerned residents have banded together in resistance, submitting letters of protest to the California Public Utilities Commission (CPUC), holding space for discussion at city supervisor meetings, and diligently signal boosting any AV incident they find. Earlier this month, activists — presumably out of fear for their safety — held a “Week of Cone,” during which they covered AV visibility sensors with traffic cones, preventing them from operation. As a result of these and other ground-up efforts, state regulators agreed to postpone the hearing on lifting restrictions until August 10th.

This is a thrilling story of grassroots, community-driven resistance successfully mobilizing in an effort to keep SF residents safe from the potentially fatal hazards that AV ridesharing represents. There’s just one problem: the story is almost entirely predicated on misrepresented statistics, contextless viral videos, and a combination of naked partisanship, conflicts of interest, and loyalty to local labor interests.

Without exception, each disturbing anecdote in the first paragraph lacks crucial context, distorts the truth, or outright lies. The small dog hit and killed by one of Waymo’s vehicles in June was off-leash, and ran out from behind a parked car at a speed that made collision almost impossible to avoid, even for the human test driver who was in the driver’s seat. A Cruise car never hindered an emergency response to a mass shooting. The car identified the emergency scene — a block party where nine people were injured in a gang-related mass shooting (which oddly received less attention from the Board of Supervisors) — performed a U-turn, and pulled over to the curb. This was explicitly corroborated by first responders. The Twitter user Jason Prado’s “near death experience” with a Waymo was a clearly benign interaction as evidenced by videos he himself posted of the incident, which depict the car yielding, then slowly driving around the man as he stands — recklessly — in the middle of the road at a crosswalk. Waymo’s response also formed the basis of the Community Note attached to his tweet: “Our car entered the intersection on a green light, and the cars ahead were backed up so we chose to appropriately yield prior to the crosswalk, ensuring you & others could safely cross.”

It’s unclear why or how the press has so completely distorted this story, but the publicly-available data is clear: at least in terms of safety, SF’s Waymo-Cruise rollout has been an unambiguous success. Both Waymo and Cruise cars have now driven over a million autonomous miles in San Francisco (1). During their first million driverless miles, Waymo vehicles were in only two reportable (2) collisions, though “experienced no collisions at all of the types that are responsible for 94 percent of fatal collisions.” Further, both of the reportable collisions were caused by human drivers (3). It should also be noted that “every vehicle-to-vehicle event involved one or more road rule violation and/or dangerous behaviors on the part of the human drivers in the other vehicle.” During the Cruise fleet’s first million driverless miles, the company reported 53 percent fewer collisions, 92 percent fewer collisions in which the driver (in this case, Cruise) was at fault, and 73 percent fewer collisions with “meaningful risk of injury,” than their human performance benchmark (4).

The most defensible — if still highly anecdotal — safety-related argument against AVs concerns not collision or direct mortality/injury, but harms that could be caused when the vehicles block traffic. In June, dozens of Cruise taxis stalled in the middle of a street for hours until a crew arrived to tow them away. And San Francisco fire chief Jeanine Nicholson recently said that AVs aren’t “ready for prime time,” citing an AV taxi that ran through emergency yellow tape, and drove half a block with wires from the scene tangled in its gear. In two more cases, AV taxis blocked firehouse driveways, thus preventing trucks from responding in a timely manner. These are real concerns. Of course, human-driven cars block first responder access all the time, which is why first responders are trained to deal with congested traffic conditions (5) — whether caused by an improperly parked AV or an improperly parked human-driven car.

But the more well-contextualized, boring truth concerning AV safety has not stopped local activists, including members of the city’s government and public agencies, from attempting to obstruct the Cruise/Waymo rollout at every turn. In a public workshop evaluating the AV rollout on June 22, a representative of the San Francisco Municipal Transportation Agency (SFMTA) presented a bar graph of AV accidents, showing a dramatic spike in 2023. The graph was picked up by local journalists, who uncritically cited the “six-fold” increase in monthly reported AV incidents. But the data wasn’t normalized for increased fleet size or expanded hours of operation. In other words, there were more AV-related incidents because there were more rides. The number of incidents was still trivial. An error in data interpretation common among high school students who have yet to take an introductory stats class.

As Lee Edwards pointed out on Twitter, according to the city’s methodology, a mass-produced car like the Toyota Camry would be the most dangerous car on the road, while a custom-built sports car would be the least dangerous.

In late June, two weeks before the CPUC’s scheduled hearing on the resolution to expand AV operations, the SFMTA, County Transportation Authority, and Mayor’s Office of Disability submitted letters of protest to the CPUC. They cited four collisions as support for their position that driverless taxis are less safe than human-driven taxis, and recommended the CPUC create new driverless readiness metrics (6), disallow AV deployment in downtown SF and during peak commuting hours, and “convene a workshop” to discuss data collection and accessibility concerns. In other words, red tape the cars to death.

But about those accidents: three of the four collisions cited by the city involved Waymos being rear-ended. In the fourth, the “collision did not include any contact” with the driverless car. The CPUC responded to these complaints by writing that “San Francisco’s arguments are not within the grounds for a proper protest,” and accused the agencies of data manipulation, citing concerns with its “statistical methods for assessing frequency of collisions and the lack of contextual awareness in assessing responsibility of the collisions cited,” and noting that the city’s “analysis appears to omit or overlook relevant facts present in the data and collision narratives that are critical for understanding the context of the cited incidents.”

City officials have taken the same line as the agencies. Jeffrey Tumlin, SFMTA Director, tweeted AVs are “creating rising traffic and safety problems with emergency vehicles, construction sites, Muni, etc. We need real safety data before unlimited expansion.” He cited an article that itself cites the content of the CPUC-rejected protest letters, and the debunked dead dog anecdote. Aaron Peskin, president of the city’s Board of Supervisors, tweeted “this is one policy area where every level of San Francisco government from [the Fire Chief] to Mayor to Muni director agree: AV taxis are not ready for prime time — and represent real safety concerns we hope CPUC takes seriously.” He cited the anti-AV Los Angeles Times article that supports its thesis with the SF fire chief’s concerns about AVs blocking traffic (discussed above), and the debunked mass shooting anecdote, failing to mention first responders say it never happened — only that if it did happen, it would have been “catastrophic.”

Given the data unambiguously shows AVs are safer than human drivers by several factors, and thus have significant life-saving potential, why are local activists, officials, and agencies in such vehement opposition? SF native and Y-Combinator president Garry Tan explains SF officials are locked in a power struggle with state-level officials, suggesting local politicians are angry California decided to regulate AVs with state-level organizations like the CPUC and DMV, which they see as encroaching on their territory. They’re fighting back by sabotaging Waymo and Cruise. But what if opposition to AVs has nothing to do with safety at all?

Organized labor stands in lock-step opposition to the AV rollout, ostensibly for safety concerns. But it’s worth noting a successful rollout of the technology would mean less human labor on the road. Given there is no current data supporting the argument that autonomous vehicles aren’t safe, might this have something to do with union concern? A direct tie to organized labor, often in terms of vital endorsements or political pressure, appears to be the most significant driving force behind AV opposition.

Fire chief Nicholson, whose position against AVs has been the basis for much of the media’s anti-AV coverage, was essentially installed by SF Firefighters Local 798, which opposes SF AVs along with other local firefighters unions. Supervisor Aaron Peskin was endorsed by the anti-AV San Francisco Taxi Workers Alliance (SFTWA) in 2016, is name-checked in a Teamsters press release calling on legislators to support AB 316, and recently posted this video advocating for Teamsters 350. Tenderloin supervisor Dean Preston, who tweeted in June, “What do driverless cars actually add to our City?” after mocking Waymo at a May supervisors meeting, is endorsed by the Teamsters and SFTWA. Jason Prado, poster of the “near-death experience” that wasn’t, is the Director of Platform at Drivers Cooperative, a worker-owned rideshare company — and direct competitor of Waymo and Cruise — who is “not interested in VC-backed or privately-owned corps unless the staff are unionized,” according to his LinkedIn.

In a more humorous example of Teamsters influence, the union appears to have astroturfed the anti-AV account @Dogs4Drivers, which tweets pictures of dogs with captions like “I can’t drive this thing by myself. Neither can @waymo or @cruise. Let’s pass AV regulation in California,” and whose followers are almost exclusively employees of the Teamsters and Berlin Rosen, a New York-based PR firm that works with labor unions.

Beyond informal ties and endorsements, most of the major anti-AV players use the exact same language: on June 8, SF Examiner published a piece called “Are Cruise’s self-driving cars ready for prime time?” On June 22, fire chief Nicholson told Los Angeles Times that AVs are “not ready for prime time”; the same day SF supervisor Aaron Peskin tweeted “AV taxis are not ready for prime time — and represent real safety concerns we hope CPUC takes seriously.” A June 29 Los Angeles times piece concludes that “what’s on the ground today is not ready for prime time.” On June 30, the Teamsters tweeted “Even more proof that ‘self-driving’ cars from @Waymo and @Cruise aren’t ready for primetime.”

Writing for the SF Chronicle this past week, columnist Soleil Ho, last seen taking up for a man with a machete at the city’s recently-shuttered Mad Max Whole Foods, dismissed concerns about AV safety in an article opposing AVs. The real problem has little to do with pedestrian safety or traffic congestion, Ho writes. The thing to really be worried about is an “autonomous car will never get tired, and it’ll certainly never try to unionize.” Here, at least Ho is being honest. It’s a shame SF elected officials and transit authorities are incapable of doing the same.

-Sanjana Friedman and Brandon Gorrell

---

1. A million miles is roughly equivalent to 40 trips around the Earth; given that the average American drives around 13,500 miles in a year, most would not drive this many miles in their lifetimes.

1. National Highway Traffic Safety Administration reports a collision if the police are called and at least one vehicle is towed away.

1. In the more severe of the two collisions, “a Waymo vehicle was hit from behind by another car whose driver was looking at their cell phone while approaching a red light.” The minor collision happened when a human driver swerved into the lane in front of the Waymo and quickly braked, such that the Waymo was unable to avoid rear-ending the car despite “exert[ing] maximum brake force.”

1. And only one of these two collisions resulted in a reported injury. In this incident, the human driver was driving over the speed limit in a turn-only lane and collided with the Cruise as it made an unprotected left turn. Police assigned responsibility to the human driver, but Cruise still chose to recall its software and code in additional defensive driving behaviors in response.

1. The U.S. Department of Transportation recommends all emergency responders train in Traffic Incident Management, a broad array of practices and procedures designed to detect, respond to, and clear traffic incidents.

1. There’s a five-year-old framework already in place.