CEI and R Street Comments to NHTSA on Automated Driving Systems 2.0: A Vision for Safety

View Full Document as PDF

On behalf of the Competitive Enterprise Institute (“CEI”) and the R Street Institute (“R Street”), we respectfully submit these comments in response to the National Highway Traffic Safety Administration’s (“NHTSA”) Request for Comments on the Automated Driving Systems: A Vision for Safety (“RFC”).[1]

CEI is a nonprofit, nonpartisan public interest organization that focuses on regulatory policy from a pro-market perspective.[2] R Street is a free-market think tank with a pragmatic approach to public policy challenges.[3] CEI and R Street previously submitted comments to NHTSA in response to its Request for Comments on the Federal Automated Vehicles Policy in September 2016.[4] CEI’s Scribner and R Street’s Adams appeared on discussion panels at NHTSA’s December 12, 2016, Federal Automated Vehicles Policy Public Meeting.[5]

Our comments are divided into the following sections, which correspond to the sections of the Automated Driving Systems guidance document subject to the RFC:

I.    Voluntary Guidance; and

II.  Technical Assistance to States.

I.Voluntary Guidance

In September 2017, NHTSA released its revised guidance document, Automated Driving Systems: A Vision for Safety 2.0 (“ADS 2.0”).[6] CEI and R Street appreciate NHTSA’s decision to eliminate the most controversial, confusing, and counterproductive elements of the 2016 Federal Automated Vehicles Policy (“FAVP”). NHTSA’s decision to repeatedly highlight in ADS 2.0 that this is a voluntary guidance document, not a binding regulation or recommendations for states to codify in any degree as a condition for permit approval, is a welcome change.[7]

In response to the 2016 FAVP’s inclusion of privacy as one of the 15 elements of the Safety Assessment Letter, we argued that NHTSA should “embrace the framework for determining when notice, consent, and disaffirmation are required that is currently employed by the Federal Trade Commission.”[8]

NHTSA’s ADS 2.0 embraces this framework by jettisoning the privacy and data sharing elements from its Voluntary Safety Self-Assessment (“VSSA”) process, recognizing that “privacy is not directly relevant to motor vehicle safety and, generally, it is the Federal Trade Commission (FTC) and not the U.S. Department of Transportation or NHTSA that is charged with protecting consumer privacy.”[9] NHTSA also notes that “there are significant concerns amongst the Industry with the sharing of proprietary intellectual property information” and “NHTSA’s current focus is on data recording needed for crash reconstruction,”[10] rather than data sharing.

We also appreciate the elimination of the FAVP’s Ethical Considerations element from ADS 2.0. As we noted in our 2016 comments, “NHTSA’s claims about ethical considerations assume vehicle automation system developers are capable of meaningfully addressing them at this time and that their attempting to do so would be socially desirable, yet fail to demonstrate that this is in fact the case.”[11]

Finally, with respect to the publication practices upon which NHTSA should rely in disseminating the submitted VSSAs to the public, we recommend the following basic framework from CEI Vice President Jim Harper:

Four key data practices that support government transparency are: authoritative sourcing, availability, machine-discoverability, and machine-readability. The first, authoritative sourcing, means producing data as near to its origination as possible—and promptly—so that the public uniformly comes to rely on the best sources of data. The second, availability, is another set of practices that ensure consistency and confidence in data.

The third transparent data practice, machine-discoverability, occurs when information is arranged so that a computer can discover the data and follow linkages among it. Machine-discoverability is produced when data is presented consistent with a host of customs about how data is identified and referenced, the naming of documents and files, the protocols for communicating data, and the organization of data within files.

The fourth transparent data practice, machine-readability, is the heart of transparency, because it allows the many meanings of data to be discovered. Machine-readable data is logically structured so that computers can automatically generate the myriad stories that the data has to tell and put it to the hundreds of uses the public would make of it in government oversight.[12]

In practice, this means NHTSA should promptly publish all submitted VSSAs in a central, searchable repository on its public website. Individual VSSAs should also be published in XML format to ensure VSSAs are not only human-readable but machine-readable, and they should be published as bulk data to permit downstream uses of VSSA information, as well. Each iteration of a submitting entity’s VSSA should be retained and remain available on NHTSA’s public website.

The VSSA database should also be structured in a manner to allow the public to easily search and sort VSSAs by characteristics such as date, manufacturing entity, and VSSA iteration. Having submitting entities register for online user accounts may ease this process and reduce the burden on NHTSA staff.

NHTSA should also allow non-submitting members of the public to create online user accounts to enable email push notifications when VSSAs are submitted or updated. The push notification interface settings should allow users to select a number of options, such as notification upon receipt of any new submitters and notification upon receipt of new VSSA iterations by a specific submitting entity. Notification triggers should be able to be set based on any searchable characteristic within the system.

Further, in addition to receiving reports based on substance strictly as they are received, users should also have the option of receiving reports of tracked activity collected during defined periods. This would allow users to review all submissions related to their areas of interest submitted over a particular period of time. Like the immediate notifications, these reports should be capable of being narrowed by any available searchable characteristics.

II.Technical Assistance to States

In NHTSA’s 2016 FAVP, CEI and R Street noted a contradiction between the nonbinding nature of the guidance document and its request that states mandate compliance with the Safety Assessment Letter as a prerequisite for entities receiving a test permit from a state regulator.[13] We warned that this “contradictory guidance is already causing confusion in the states,”[14] as in the days following the release of the FAVP, the California Department of Motor Vehicles released draft rules requiring manufacturers to “certify that testing will be conducted in accordance with the National Highway Traffic Safety Administration’s ‘Vehicle Performance Guidance for Automated Vehicles’”[15] as a prerequisite for obtaining a testing permit.

We urged NHTSA to “make clear that it is inappropriate for states to attempt to mandate compliance with a non-binding federal guidance document” in its revision to the FAVP.[16] In ADS 2.0, NHTSA wisely eliminates this contradiction while also highlighting in clear, explicit terms that “NHTSA strongly encourages States not to codify this Voluntary Guidance (that is, incorporate it into State statutes) as a legal requirement for any phases of development, testing, or deployment of ADSs.”[17]

We believe the updated guidance to states clearly reflects the respective roles of federal and state legal authorities and administrative competencies, while appropriately cautioning states against attempting to occupy NHTSA’s regulatory field in the absence of Federal Motor Vehicle Safety Standards that directly address automated driving systems.

Finally, we urge NHTSA to revise the ADS 2.0 section on Technical Assistance to States to include a recommendation that states include a driver license reciprocity provision in any testing regulations, as we did in our comments on the 2016 FAVP.[18] The FAVP stated, “The operators testing the vehicles must hold a valid state driver’s license.”[19] As we noted, “This could be reasonably interpreted by states that potential test drivers should possess a valid driver license issued by the testing state”[20] and that “[t]his unnecessarily restricts the potential test driver labor pool while providing no discernable safety benefits, and poses additional problems for developers wishing to test their vehicle automation systems in metropolitan areas that span across state lines, such as Washington, D.C., New York City, and St. Louis.”[21]

ADS 2.0 improves on the FAVP by stating that “[s]tates could request” the “[i]dentification of each test operator, the operator’s driver license number, and the State or country in which the operator is licensed.”[22] NHTSA should add a sentence explicitly encouraging states to accept a test operator’s driver license, regardless of the issuing state or country.



[1].     Request for Comments on the Automated Driving Systems: A Vision for Safety, Notice, NHTSA-2017-0082, 82 Fed. Reg. 43321 (Sep. 15, 2017).

[2].     See About CEI, https://cei.org/about-cei (last visited Oct. 20, 2017).

[3].     See About R Street, http://www.rstreet.org/about/why-r-street/ (last visited Oct. 20, 2017).

[4].     Comments of the Competitive Enterprise Institute, R Street Institute, & TechFreedom on the Federal Automated Vehicles Policy, Notice, NHTSA-2016-0090, 81 Fed. Reg. 65703 (Sep. 23, 2016), available at https://www.regulations.gov/document?D=NHTSA-2016-0090-1000 [hereinafter Comments of CEI et al.].

[5].     Transcript of the National Highway Traffic Safety Administration’s Federal Automated Vehicles Policy Public Meeting, Arlington, Va. (Dec. 12, 2016), available at https://www.regulations.gov/document?D=NHTSA-2016-0090-1130.

[6].     National Highway Traffic Safety Administration, “Automated Driving Systems: A Vision for Safety 2.0,” U.S. Department of Transportation (Sep. 12, 2017), available at https://www.nhtsa.gov/av [hereinafter ADS 2.0].

[7].     Id.  at 12.

[8].     Comments of CEI et al., supra note 4, at 5.

[9].     National Highway Traffic Safety Administration, “Manufacturers: Automated Driving Systems,” U.S. Department of Transportation, available at https://www.nhtsa.gov/manufacturers/automated-driving-systems (last visited Oct. 20, 2017).

[10].    Id.

[11].    Comments of CEI et al., supra note 4, at 9.

[12].    Jim Harper, “Publication Practices for Transparent Government,” Cato Institute Briefing Paper No. 121 (Sep. 23, 2011), available at https://object.cato.org/sites/cato.org/files/pubs/pdf/bp121.pdf.

[13].    Comments of CEI et al., supra note 4, at 11.

[14].    Id. at 12.

[15].    California Department of Motor Vehicles, “Revised Draft Deployment Regulations: Autonomous Vehicles Express Terms,” § 227.4(d) at 3 (Sep. 30, 2016), available at https://www.dmv.ca.gov/portal/wcm/connect/211897ae-c58a-4f28-a2b7-03cbe213e51d/avexpressterms_93016.pdf?MOD=AJPERES.

[16].    Comments of CEI et al., supra note 4, at 12.

[17].    ADS 2.0, supra note 6, at 18.

[18].    Comments of CEI et al., supra note 4, at 12.

[19].    National Highway Traffic Safety Administration, “Federal Automated Vehicles Policy,” U.S. Department of Transportation, at 43 (Sep. 20, 2016), available at https://www.nhtsa.gov/sites/nhtsa.dot.gov/files/documents/13069a-ads2.0_090617_v9a_tag.pdf.

[20].    Comments of CEI et al., supra note 4, at 12.

[21].    Id.

[22].    ADS 2.0, supra note 6, at 23.