Creating adaptable AV policy to match AV technology

In the absence of federal standards, states have started trying to solve the knotty policy and regulatory challenge of taking humans out of the driver’s seat. Since we know self-driving technology will keep evolving, any policy framework should set reasonable performance expectations for technology while building in the flexibility to adapt to innovation. But it can’t be so flexible that it amounts to no framework at all.

The promotion of a national testing strategy also helps AVs deliver maximum anticipated social benefits, like mobility for the disabled, for aging residents, and for areas whose economic and geographic limitations leave them underserved.

We recently helped the State of Vermont finalize a framework for AV testing and write accompanying guidelines, a project that excited us for several reasons. Vermont does not have perfect roads—the State counts 7,151 miles of paved roads, 6,008 miles of gravel roads, and 1,590 miles of graded earth roads. Vermont has challenging weather, with between 5 and 10 feet of snowfall per year interspersed with freezing rain. Vermont has aging and rural populations—27% of the total population is over 60 years old and 65% of the population lives in rural communities. And, yes, Vermont even has a mud season. Taken together, these challenging conditions make Vermont a place where testing should take place to ensure that AVs can operate successfully and reliably.

In developing Vermont’s framework, a snapshot of issues emerged, which many states and communities are confronting:

• Who can regulate safety? More candidly, who verifies the claims that AVs being tested on public roads qualify under Federal Motor Vehicles Safety Standards? How existing roles around licensing and registration (think human driving tests) reconcile with federal jurisdiction over vehicle safety that will be controlled by onboard computer systems is yet to be determined. (Note: In late 2020, the National Highway Traffic Safety Administration  announced that it would begin developing a framework to govern automated-driving-system safety.)
• How to modernize insurance and liability partnership models? Wider testing of self-driving vehicles gives greater urgency to the task of devising new liability-and-risk-mitigation models designed to share risk. A partnership to promote an evolving technology like AVs requires an investment of time and resources from both private and public sides, and it provides an opportunity for new insurance models that share risk proportionately to encourage more testing.
• What is the right level of data reporting? Public agencies rightfully want access to data about the operation of AVs on public roads, but the private sector expresses strong concern about privacy, data security, and trade secrets. New policies should promote alignment around goals and expectations at the outset while also accounting for open-records laws. (We recently helped the Transportation Research Board produce a white paper on this topic.)
• What is the right amount of interaction with law enforcement and local coordination? While local coordination can look like a barrier to testing and deploying AVs, it can also be viewed as an opportunity to educate, build trust, and promote adoption of AVs. Local coordination also can make sure the integration of AVs aligns with transportation planning documents and identify use cases that may center around tourism opportunities.  

By addressing such complicated questions in our work with the Vermont Agency of Transportation, the following approaches emerge for consideration to support a national testing and deployment strategy for AVs:

• Take a mode-neutral approach. AVs won’t just carry passengers, they’ll deliver goods and services, a fact that the COVID-19 pandemic has brought into sharp focus. Policy frameworks, including access to right-of-way, fee structures, and safety considerations, will need to distinguish between passenger and goods delivery (which may never have humans in a vehicle and may be smaller and lighter from a safety regulatory perspective).
• Set standards that balance safety and innovation while technology evolves and we identify use cases for AVs. Can we, for example, expand existing standards built around human drivers to accommodate AVs? Or should we build a new regulatory structure from scratch to cover a range of next-generation mobility solutions already being developed and tested?
• Use the need for AV policy to tackle larger issues—like the long-discussed need to rebuild infrastructure. Next-generation mobility solutions like self-driving vehicles present an opportunity to break down silos and tackle this crucial issue. Investment in a broad approach to planning, redesigning, and inspiring people-focused infrastructure can maximize AVs’ anticipated social benefits. It can also spur investment in a transportation network that gains resilience from a broader set of more balanced choices.
• Promote secure and responsible data sharing. We’ll need frameworks that align goals and expectations, ensure privacy and data security, and use anonymized data to plan, inform decision-makers, and ensure public education to support investment in the future of mobility. This starts with setting up a task force to promote discussions around goals and proposed uses of data, while making sure data collection incorporates best practices.

Without smarter and more adaptable policy and regulatory structures, we could easily lose some of the social benefits of AVs. This fluid (yet foundation-setting) moment calls for collaboration across all sectors to maximize this opportunity to improve mobility for both passengers and goods. Through a national testing strategy that promotes public and private partnerships, we can encourage investment and strengthen confidence in this emerging technology.