In 2017, the U.S. Senate introduced the American Vision for Safer Transportation through Advancement of Revolutionary Technologies Act (the “AV START Act”). The main purpose of the Act is to establish an interim framework to facilitate the deployment of autonomous vehicles on U.S. roads and to permit the sale or commercialization of highly autonomous vehicles despite two key obstacles.
The first obstacle is that vehicles with automated driving systems cannot be certified as complying with some existing federal motor vehicle safety standards (for example, those predicated on a human driving the vehicle). The second is that there are no safety standards applicable to autonomous-vehicle design and performance. The Act aims to facilitate the ongoing development and deployment of autonomous-vehicle technologies while safety standards are developed.
To achieve this aim, the Act establishes the following framework:
- Identify existing safety standards and test procedures that apply to or are predicated on a driver being a human and figure out how to revise them to accommodate an automated driving system as the driver.
- Conduct rulemaking to revise existing standards and test procedures so they accommodate and permit compliance by automated driving systems. Separately, a technical committee is to be established to recommend, within five years, new standards that are necessary and appropriate to ensure automated-driving-system and highly automated-vehicle safety. NHTSA then would have a year to evaluate the recommendations and commence rulemaking for adopting those standards.
- Amend the existing statute that allows for temporary exemptions that permit, on approval of NHTSA, the sale or introduction into interstate commerce of a limited number of a manufacturer’s vehicles for purposes of evaluating new motor vehicle safety features that do not comply with safety standards. Under the amendment, if the safety feature is one that enables a highly automated vehicle or automated driving system, the manufacturer may be authorized to sell or introduce into interstate commerce up to 100,000 vehicles per year until new standards permit such features.
- Because autonomous vehicle-related safety standards won’t exist for some time, establish an alternative short-term process for establishing that highly automated vehicle technologies are safe. That process requires the submission by a manufacturer of a safety evaluation report (previously referred to as safety self-assessments) explaining how it considered and addressed certain safety-related issues or subjects relating to the design, function and operational characteristics of its automated driving system or highly automated vehicle before the vehicle is tested, sold or otherwise commercialized. Under existing NHTSA guidance, manufacturers are only “encouraged” to publish similar reports to the public.
START Stops in Senate
After the AV START Act breezed through committee, it stalled in the full Senate due to the objections of a few senators. Their objections included:
- The interim framework does not result in rapid enough development of new AV-related safety standards, and it should provide the same level of safety to vehicle occupants in a crash as required in current standards.
- Concern that the bill indefinitely pre-empts state and local regulations even if federal safety standards are never developed. They request a sunset on interim pre-emption.
- The bill does not do enough to develop cybersecurity safeguards or protect consumer privacy.
- The requirement for submitting safety-evaluation reports also should apply to Level 1 and 2 AVs (citing the two publicized Tesla crashes in Florida and California), not just highly automated vehicles.
- Underlying these objections appears to be a concern that the technology is still, relatively speaking, in its infancy and not yet proven completely capable or reliable. Those who hold this view believe the framework does not assure the same level of safety as would new autonomous-vehicle safety standards. To address these concerns while also preserving the framework, proponents may want to consider some revisions of the bill.
With respect to temporary exemptions allowing the sale of autonomous vehicles, the Act amends an existing statute that allows NHTSA to grant an exemption allowing the sale of up to 2,500 vehicles for evaluating safety features that don’t comply with existing standards but which provide the same level of safety as vehicles that are compliant.
It can take NHTSA months just to publish notice of receipt of a petition for temporary exemption. It can take up to two years for it to review and rule on an exemption, and that is for exemptions to permit the sale of vehicles with things like a new type of headlight, not autonomous vehicles lacking steering wheels and pedals.
Under the Act, NHTSA will have only 180 days to grant or deny a petition for an exemption. And the exemption can be for up to 100,000 vehicles. Thus, the bill constrains NHTSA’s review of requests for exemptions that, if granted, will place a large number of previously unsaleable vehicles with new technologies on the road.
With respect to required submittal of a safety evaluation report intended to show that highly automated vehicles are safe, NHTSA’s prior comments about virtually identical safety self-assessments raises questions about whether the reports will accomplish what is intended. As indicated, in its Guidance 2.0, NHTSA considered such reports to be promotional pieces aimed at the public, not a rigorous technical report intended for submittal to NHTSA. Safety assessments that NHTSA viewed as promotional pieces are, under the bill, the principal method of assuring the safety of highly automated vehicles.
To be fair, the description of the safety evaluation report in the Act includes a requirement that manufacturers address, through a “documented assessment, testing and validation process,” each of the specified subject areas, which largely mirror the safety elements in Guidance 2.0. The documentation requirement does make the process more rigorous, but the nature of the “documentation” is not set forth and the Act prohibits NHTSA from “conditioning” the testing or sale of highly automated vehicles on “a review” of the contents of the reports.
Also, the report need only be filed upon testing and 90 days before the vehicle is sold, leaving little time for review.
While portions of the legislation may be seen as problematic, completely abandoning this interim framework in favor of the existing process of establishing autonomous vehicle-related safety standards and requiring manufacturers to self-certify their autonomous vehicles is not preferable because that would effectively delay the introduction of this important technology for 5-10 years. However, to push the bill over the line, the industry, regulators and the legislature should consider making some changes to the proposed law to make the analysis of the safety of autonomous vehicles somewhat more rigorous.
A start would be to make NHTSA’s review of the safety evaluation reports less perfunctory. The law could grant NHTSA sufficient time to review the reports (and petitions for exemptions). It could permit NHTSA to condition the testing and sale of highly autonomous vehicles on the completeness of the reports. Along these lines, it could set forth minimum documentation requirements that support the performance and safety claims in the reports and allow NHTSA to reject reports that lack sufficient documentation or are generic rather than vehicle-specific.
Finally, autonomous vehicles have been under development for years and industry standards for autonomous vehicle technologies already exist. The timetable in the bill for establishing new safety standards could be advanced.
Richard Wilhelm is Of Counsel at Dickinson Wright PLLC’s Detroit office, where he focuses his practice on automotive law and product liability and personal injury. [email protected].