For over a decade, we were assured by tech leaders that self-driving cars were coming “soon.” Following years of hype, they appear to have finally delivered. In San Francisco, autonomously driven Waymos can be seen roaming the streets; one can book a Waymo as easily as one would an Uber. And in Philadelphia, following a round of tests with safety drivers, Waymo is experimenting with fully autonomous operation of its fleet in preparation for a future public launch.
Nationwide, this rollout has been less than perfect. Late last year, a car operated by Waymo in Los Angeles breezed through a police stop. Another incident in Boston saw a Waymo run over a cat. It is events like these that have left the door open to criticism; all sides of the political spectrum are coming to weigh in. Conservative pundit Matt Walsh recently declared on his podcast that “I would rather be killed because my Pakistani driver named Saeed Muhammad or whatever drove straight into oncoming traffic. […] over a soulless robot self-driving car malfunctioning.” During a Boston City Council hearing, an aggrieved councilor decried the language of a Waymo representative who used the term “Waymo driver” in reference to the car. “Drivers are people,” she declared “I think that is so creepy, and I think that is so unsettling.”
Yet, reducing this to an ontological debate over what a “driver” is (or whether an accident is preferable because it is human-caused) risks ignoring the far more unsettling record of human drivers. It is grim. Last quarter—on Drexel University’s campus alone—two people died in hit-and-runs. One on 33rd and Market (where as of writing this, no arrests have been made) and the other on 36th and Market. This happens hundreds of times each day nationwide. Unlike the mistakes made by AVs, human-caused car accidents seem to hardly ever register, even on local news. This is especially jarring given that recent studies seem to indicate the relative safety of AVs compared to human drivers, as do recent reports by insurance companies.
Though I count myself among the many Americans skeptical of the merits of the rapid rollout of AI—which is displacing labor and is set to trigger an economic bubble—the assertion that an infinite slop machine has much to do with self-driving cars—however similar the underlying technology—is dubious at best. That companies use AI to degrade services around us should not cause us to avoid a demonstrably lifesaving technology. Especially given that because AVs collect data used to train as they drive, today’s AVs are the worst they will ever be. Their deployment would mean lower traffic fatalities now and in the future. It is especially surprising to see progressives make arguments against this that read almost conservative in their effect. There is nothing progressive in simply declaring a better world to be impossible.
Other concerns address transit accessibility. The Urban Institute, for instance, claimed in 2018 that “access to self-driving cars might reduce demand for public transit” which would in turn “increase overall transit costs for low-income households.” But because SEPTA is seen as unsafe and unreliable with many residents having switched to using rideshare apps, the demand for public transit is already “low.” It is these people—not residents who continue to ride with SEPTA—that are likely to start using AVs if given the option, leaving public transit ridership ultimately unaffected. While the loss of state funding has left SEPTA’s budget in crisis, and this has led to fare hikes, I cannot find coherence in a logic that demands withholding AV deployment until we solve this.
Another point often brought up is the loss of rideshare jobs. Setting aside the fact that the rideshare industry itself was created by displacing taxi drivers, it is an industry that—accounting for car depreciation—often pays its workers below minimum wage. Crucially, both drivers and passengers in rideshare vehicles (especially women) often report feeling unsafe. We should direct public policy to create higher-paying, skilled work as opposed to protecting the current state of the gig economy.
In fact, the relevant question is not whether technology displaces labor—it always will—but whether preventing deaths justifies that displacement. As then-Governor Tom Wolf declared when he signed HB 2398 to permit AVs to drive on Pennsylvania’s roads: “often new technology brings job replacement, and we must ensure the Pennsylvania workers are protected and allowed the opportunity to participate in this industry.” While Harrisburg works on the details of this new regulatory framework, Philadelphia would do well to demand rigorous safety standards for AVs, rather than oppose their deployment entirely. Because, as Nidhi Kalra writes “If we wait until these vehicles are nearly perfect, our research suggests the cost will be many thousands of needless vehicle crash deaths caused by human mistakes. It’s the very definition of perfect being the enemy of good.”