Trust the Auto Industry with Automated Vehicle Safety? Really?

Shutterstock photo by Anastasiya Aleksandrenko

How much leeway should the industry that brought us Unsafe at Any Speed, the Ford Pinto, and, most recently, the General Motors ignition switch scandal be given in assuring the safety of the automated driving systems it hopes to build into the vehicles Americans drive?

Quite a bit, if you listen to the Trump administration and Congress. The U.S. House overwhelmingly adopted legislation to allow manufacturers to place as many as 100,000 automated vehicles a year per company on the roads within several years. Key leaders in the U.S. Senate reached agreement on text of a parallel bill last week. In September, the Trump administration released its regulatory guidance for automated driving systems, a document chockablock with requests that entities involved in automated vehicle development “consider” this or “be encouraged to” do that, but very little in the way of clear standards for protecting the safety of drivers and other road users.

We have long argued that autonomous vehicles can play an important role in the transition to a more sustainable, safer and saner transportation system – particularly if they expand availability of shared mobility services that reduce the innate incentives to drive that come with vehicle ownership, free up valuable urban land, and enable the “rightsizing” of the vehicle fleet.

Automakers argue that a too-aggressive approach to regulation could slow or stifle that transition. They have a point: testing in real-world conditions is essential for further development of the technology, and a poorly considered or overzealous “patchwork” of regulations across the 50 states could hamstring the roll-out of technologies with potential to save lives.

But the history of the automobile industry suggests that a hands-off regulatory approach brings its own dangers – and could potentially limit America’s ability to obtain the full benefits from the adoption of autonomous vehicles.

The transition to automated vehicles has been promoted – rightly – as an opportunity to achieve a once-in-a-generation step-change improvement in road safety, delivering a drastic reduction in the crashes that claimed more than 40,000 lives and seriously injured 4.6 million Americans last year. Unfortunately, the standard in the House bill for automakers to receive a safety waiver for autonomous vehicle testing merely requires that vehicles be safer than those driven by humans. As someone who bikes, walks and drives in Boston, I can tell you that the standard of “better than a human driver” is a pretty low bar to meet. We need to shoot higher – much higher – in our demands for better safety if the autonomous vehicle revolution is to pay dividends.

Similarly, advocates of sustainable transportation and livable cities see autonomous vehicles as an opportunity for a phase shift in how America’s transportation system operates. Taking advantage of that potential requires cities and states – the levels of government that actually manage the use of vehicles on the roads – to have a role in their design, regulation and evaluation from Day One. The House bill has been criticized by Transportation for America and the National Association of City Transportation Officials for failing to provide cities and states with the information they need to evaluate how AVs operate in real-world cities, and for limiting the ability of cities, states and federal officials to intervene in situations when safety is compromised. As Henry Grabar writes in Slate:

Stakeholders at the state level, including the National Conference of State Legislatures and the National Governors Association, say the legislation that passed the House appears to shift even the regulation of vehicle operations (i.e., speed limits and idling laws), not just design and safety standards, from states to Washington. (When operations are coded into the cars in the factory, where do you draw the line between design and actual behavior on the road?) It’s not clear whether the law would permit states to, for example, tax or outlaw the use of autonomous vehicles without passengers, which could clog up city streets.

A more basic question at play is whether the auto industry (along with the companies seeking to provide the technologies used in highly automated vehicles) can be trusted to use any regulatory flexibility they are given to maximize vehicle safety.

Technological advances do not necessarily lead to net improvements in safety. While the technological revolution of the past decade has yielded important safety advances in vehicles – from rearview cameras to lane departure warnings – it has also allowed automakers to build distracting entertainment and information technology into vehicles that potentially puts road users at risk.

The AAA Foundation for Traffic Safety has produced a series of cutting-edge reports measuring the degree of distraction these systems create and their impacts on traffic safety, including a 2015 study that found that in-vehicle information systems create moderate to high levels of cognitive distraction for users. The study found significant differences in the cognitive workload of various systems, with the most distracting of them requiring more attention than conducting a cellphone conversation (already a dangerous level of distraction). Moreover, drivers remained impaired for a full 27 seconds, on average, after they completed their interaction with the systems.

At the same time, automakers continue to make vehicles designed to be misused and to advertise them in ways that encourage and normalize unsafe behavior on the roads. The Automotive News earlier this year called for the 840 horsepower (!) Dodge Challenger SRT Demon to be banned from public roads, even though it meets federal safety standards, because of its propensity to be misused. A 2005 study found that 45 percent of all North American auto advertisements included an unsafe driving sequence, with 85 percent of those sequences including aggressive driving.

These are the folks we trust to experiment with autonomous vehicles on our roads, with minimal accountability?

Let’s say that, for the sake of expediting the development of autonomous vehicles, we do need to trust the automakers to some degree – to accept the possibility of failure and mistakes in the process of learning what we need to know to achieve a vastly safer transportation system in the future.

If that is the case, the bare minimum the public should expect of the automobile industry is adoption of the Ronald Reagan Rule: trust, but verify. The flexibility to experiment with autonomous vehicles on public roads should be accompanied with transparency about how and where those vehicles operate and prompt disclosure of any safety issues that arise in their operation. Regulators – whether at the federal or state level or both – also need to be empowered to act on that information right away in order to protect the public.

The rollout of automated vehicles is an exciting opportunity to address many of the problems of our car-dependent transportation system – from carnage on the roads to global warming. But we will likely only get one chance to get it right.  The continued engagement of all levels of government will be needed to achieve a positive result. Handing over the keys to the automakers hasn’t worked to maximize safety before, and it likely won’t work this time either.