The arguments made by the autonomous vehicle (AV) industry for limited regulation and rapid deployment of self-driving car technology bear a striking resemblance to the discussion about the US regulatory response to Covid-19 vaccines. Both appeal to exigent circumstances as a reason for rapid action to save lives, even if that means throwing some precautions to the wind. It is a justification that relies on a narrative of exceptions.
The AV industry discusses AV safety against the backdrop of the language of disease, implying that the carnage on US highways is an epidemic. A review of resources listed under the topic “Social-Impacts” by Partners for Automated Vehicle Education (PAVE), the AV industry’s 501(c)(3), illustrates this tactic, posting many materials from various health experts and the Centers for Disease Control and Prevention (CDC). Academics with degrees in public health serve on PAVE’s academic advisory council, inserting the AV industry into discussions about health policy and management.
A utilitarian argument follows, noting that delayed deployment of self-driving car technology will result in a greater loss of life than a more rapid deployment: delay equals needless deaths, a theme echoed by some commentary on the AV industry. The CEO of Aurora Innovation, an AV company in the process of “going public,” noted: “there’s an incredible opportunity to move from the status quo towards zero. We should be saving those lives along the way and not wait for the perfect at the expense of all those lives.” Subsidiary arguments burnish the public health crisis narrative, focusing on transportation for the elderly and mobility impaired, as well as adverse health consequences from greenhouse gases, all listed in the PAVE resources, though the impact of AVs on greenhouse gases is uncertain.
Both regulation and public skepticism risk slowing down the deployment process. The AV industry’s solution is to convince the US government to limit regulation and to “educate” the public about the benefits of self-driving car technology, essentially trusting that the AV industry will self-regulate and “do the right thing.” The less-publicized corollary to this approach is that proceeding rapidly likely entails taking increased risks. Compare this to the discussion about how the FDA handled final approval of a Covid-19 vaccine.
In an August 24, 2021, essay in the New York Times morning newsletter, F.D.A., Not F.D.R.: One chose bureaucratic caution. One didn’t., David Leonhardt observed that the “F.D.A.’s full approval of Pfizer’s Covid-19 vaccine is already making a difference.” But he also notes that the months-long delay “has had large costs.” As an illustration of these costs, he notes that in the week before vaccine approval, approximately 1,000 Americans per day have died of Covid and that the vaccine would have probably prevented 95% of these fatalities. This argument mirrors the AV industry’s observation that approximately 94% of traffic fatalities result from human driver errors which self-driving car technology might eliminate.
The increased risks of premature approval of a vaccine are the potential side effects that might cause injury or death in some patients. The increased risk of early deployment of self-driving car technology is the potential separate risks a machine driver poses to those on our highways that went unidentified in a rapid deployment process. The task requires sufficient time to identify (and reduce) unknown unacceptable residual risks. The criticism of the FDA’s delay of full approval is that it “was more about the process than science.” The question about delay in regulatory assessment of the safety of self-driving car technology is whether the delay also is about process or science.
Apparently, the FDA had made up its mind on the substantive question of vaccine safety over three months ago when the acting FDA Commissioner stated that the vaccines “have met our high standards for quality, safety and effectiveness.” If the FDA had granted full approval earlier, the Pentagon, and some universities and large companies, would have had grounds to announce new vaccine mandates earlier. So, if the delay amounted to red-tape or foot-dragging, the criticism might be in order, and indeed some additional lives might have been lost through delay.
The New York Times essay contrasts the approach of the FDA which led to delay with the approaches of President Roosevelt to fight the Great Depression and WWII, and Ben Bernanke to address the financial crisis of 2007-2009:
In each of the instances, officials avoided taking steps that clearly violated the law. Yet they recognized that the law often includes gray areas and gives government agencies leeway to choose one of several approaches. During normal times, taking the cautious route and following procedural precedent tends to make sense. It minimizes chaos and mistakes.
But a national emergency can change the equation.
In the case of an emergency, F.D.R.’s approach was: “Above all, try something.”
The AV industry clearly wants to promote greater risk-taking for the deployment of self-driving car technology by creating the sense of an emergency akin to a health crisis like Covid-19. Emergencies justify exceptions to business as usual. The AV industry also concerns itself with other narratives of emergency to further the case for an exception, such as a wartime footing, to hasten deployment of self-driving car technology, by raising the spectre of national security.
For example, a recent report from the Center for Strategic and International Studies (funded by Cruise LLC, an AV Company based in San Francisco) stresses the pressing need for the US to take a leadership role in the development of autonomous vehicle technology, both for application to military technologies and more generally for “global influence, prestige, and power.” Significantly, the report helpfully notes that “Companies will also need a supportive regulatory environment to be able to deploy AVs at commercial scale.” The report makes clear that we are in an economic “war” with China, and that it is critical for national security that the US win this war.
The AV companies have adopted the narrative of health emergencies and war to make a case for less regulation because there is general agreement that our government ought to be less cautious in exigent circumstances. But the case of self-driving car safety is different from the case of Covid-19 vaccine safety in important respects.
First, the jury is still out on whether a machine driver is, or will ever be, safer than a human driver. No fully autonomous vehicles currently exist, with vehicles sold today rating a Level 1 or 2 on the SAE’s autonomy capability scale. So there is no immediate emergency need for special consideration to deploy AVs at scale in the very near term. In contrast, the Covid-19 pandemic was upon us.
Second, the FDA already has safety standards. In contrast, the National Transportation Highway Safety Administration has not yet promulgated any regulations governing safety for fully self-driving cars, nor has Congress passed any legislation.
Third, while the FDA presumably had a scientific basis for its earlier assessment that the vaccines met high standards for quality, safety and effectiveness, currently there is no agreement upon any standards to assess the safety of self-driving cars within either the AV industry or the regulators. As of now, there is no agreed basis to make a relative safety claim, or even make a reasoned estimate about trade-offs. An industry group only recently issued a proposed “best practice” for AV safety assessment. And, an AV company announced the first self-driving “Safety Case Framework” that applies to autonomous trucks and passenger vehicles on August 19, 2021. Currently, the NHTSA allows AV companies to self-report on a voluntary basis. The question of standards is in flux.
Fourth, no one was questioning whether or not the FDA ought to have standards for approval of medicines and vaccines or whether self-regulation would be the better approach for pharmaceuticals. Yet, the AV industry argues for self-regulation, claiming that any regulations will be outdated by the time they become effective and will impede innovation. Currently, the NHTSA posts voluntary safety self-assessments from AV companies.
Fifth, in the case of Covid-19, there really was not any practical alternative to a vaccine. Social distancing, masks and ventilators were/are a very poor substitute. In contrast, even if one thought the number of highway deaths was an epidemic, there are many other effective alternative remedies that are both proven and cost-effective (e.g. lower speed limits), whereas AV technology is currently uncertain and aspirational. As one commentator remarked, “the immediate priority must be to advance verified technology that will save lives now.”
Consideration of the differences between the FDA approval process and the state of (non) regulation of self-driving cars, suggests that the NHTSA must first, at least, announce some safety standards for self-driving cars—such as safer than a human driver at the time of deployment—before it is even proper to consider whether exigent circumstances justify speeding up an approval process, or relaxing a safety standard (as was done in the case of FDA emergency authorization for the Pfizer, Moderna and Johnson & Johnson vaccines).
One reason, however, to create a sense of urgency behind the deployment of self-driving car technology is the fear that, when the time comes to deploy at scale, neither the AV industry nor the regulators will have any standards and metrics, or sufficient data, to make a statistically significant safety assessment to a reasonable degree of scientific certainty. This is very different from the case of the Covid-19 vaccines for which some human trials had been conducted. It is particularly true that, in a state of ignorance, the emergency situation provides a rationale to do something where otherwise no rationale would exist as even a fig leaf for the regulators to hang their hats upon. If the early deployment of self-driving car technology does not go well, the regulators might parrot the words of F.D.R. as justification: “Above all, we as regulators, at least tried something.”
William H. Widen is a Professor at University of Miami School of Law.
Suggested citation: William H. Widen, Autonomous Vehicles, COVID Vaccines & The Narrative of Exceptions, JURIST – Academic Commentary, September 1, 2021, https://www.jurist.org/commentary/2021/09/william-widen-autonomous-vehicles-covid-vaccines/.
This article was prepared for publication by Giri Aravind, a JURIST staff editor. Please direct any questions or comments to him at firstname.lastname@example.org