U.S. flag

An official website of the United States government, Department of Justice.

Autonomous Vehicles: Expert Panel Lists Top Needs for Law Enforcement as the Dawn of Driverless Roads Inches Closer

A new age of algorithms taking the wheel en masse is still some years away, but law enforcement must prepare now, an NIJ-sponsored group concludes.
Date Published
February 16, 2021

The autonomous vehicle revolution will reinvent the ways that people and goods are moved, cities and roads are planned, and transportation resources are deployed and conserved.

The question is how soon — not if — driverless vehicles become common sights in driveways and rear-view mirrors. Already, across the globe, autonomous cars, trucks, street-legal delivery carts, and people-movers are popping up in limited real-life applications in communities, not just on test tracks.

Law enforcement, meanwhile, faces its own race to be ready for driverless vehicles, as it plays catch-up with deep-pocket automotive and artificial intelligence industries. The challenges of crime-stopping, traffic control, public safety, and cybersecurity on the road can only grow as more human drivers give way to algorithms.  

Consider some scenarios:

A hacker from a hostile regime is ready to electronically commandeer an autonomous armored car full of gold bars bound for a Federal Reserve Bank. What’s to stop him?

An autonomous bus rolls up to a patrol officer who is motioning to divert traffic around downed power lines? How could a driverless bus comply?

A panel of experts sponsored by the National Institute of Justice convened in 2019 to begin identifying the universe of critical law enforcement issues and needs related to the coming age of autonomous vehicles. Supported by researchers from RAND Corporation (RAND) and the Police Executive Research Forum (PERF), the panel of law enforcement, industry, and traffic safety leaders was tasked with articulating key public safety challenges presented by autonomous road vehicle development over the next five years, beginning in 2019. Panel members represented industry, law enforcement, and the research community. The panel issued its report this year.

The panel identified 17 high-priority needs (i.e., an opportunity or problem and related solution) associated with key systemic issues or challenges.

Highlights of the panel’s findings included:

Cybersecurity and Vehicle Communication Needs

  • Research on systems to enable law enforcement to identify a vehicle’s authorization to run in automated mode.
  • Research on technology that enables law enforcement to communicate with vehicles in automated mode.

Stakeholder Communication Needs

  • Workshops or ride-alongs to educate law enforcement and other agency staff on how driverless vehicles function.
  • Surveys to identify the most useful data the autonomous vehicle industry can make available to law enforcement for investigations of crashes and other incidents.

Standardization Needs

  • Model training and guidelines for interacting with autonomous vehicles running in automated mode.
  • Development of descriptions of standard behaviors (such as pulling off the road in a safe spot) that law enforcement will expect autonomous vehicles to perform across the United States.

Issue Background for Priority Needs Development

Timeframe for Development and Deployment

A threshold inquiry for the panel was setting a realistic time horizon for the development of autonomous vehicles. Although a number of early industry forecasts had autonomous fleets operating on U.S. roads by early in this decade, the experts scaled back autonomous vehicle expectations and resultant timeframe of law enforcement’s response. Within a five-year window after the study, only limited fleets of highly autonomous vehicles — such as ride-share and shuttle programs — were likely to appear, and only in a few American cities, the participants predicted. Few of those fleet vehicles would likely operate without an onboard human who could take control if or when needed. According to the project team’s report, personal use of vehicles requiring no driver at all would be very limited within that five-year window. 

That said, the panel noted, today’s cars and trucks already contain a number of autonomous features. 

Although the panel members predicted that the vast majority of law enforcement agencies will not be addressing autonomous vehicles on the road before 2024, they urged all stakeholder entities to use that time well by broadly and proactively engaging on the challenges and opportunities autonomous vehicles will pose for law enforcement.

Communication With Autonomous Vehicles

For law enforcement, the importance of being able to interface with autonomous vehicles is self-evident. One identified key need is a means of determining, first, whether any vehicle on the road is operating autonomously and, if so, at what level of autonomy, panelists observed. Determining the presence and level of engagement of a human driver will inform important law enforcement decisions, such as the procedures for making a traffic stop and potential culpability for driving behavior, the report said. Participants advocated for research on developing a standard electronic means for law enforcement to communicate securely with autonomous vehicles on the road.  

Collaboration Between Law Enforcement, Industry, and Communities

The panelists stressed the need for proactive problem solving with law enforcement, autonomous vehicle manufacturers and operators, and communities all working together. Law enforcement would benefit from having a better understanding of autonomous vehicle capabilities, and manufacturers would benefit from insights on the law enforcement implications of autonomous vehicles operating in communities, they said. Participants also saw opportunities to use data generated by autonomous vehicles — for example, videos from an autonomous vehicle passing an active crime scene — to support public safety and crime investigations. It was noted, however, that legal protections of privacy and intellectual property must be accommodated.

Evolution of Law Enforcement Training and Resources

Law enforcement will need to adapt its processes to an autonomous vehicle environment, panelists said. Crash investigations must evolve as human driving mistakes and misconduct, such as driving under the influence, are expected to become less prevalent. At the same time, criminals can be expected to exploit autonomous vehicles for their own purposes, such as electronic attacks on vehicles with ransomware, using autonomous vehicles to traffic drugs and people, and using autonomous vehicles for privacy invasion.

Panel Purpose and Process 

RAND and PERF organized the workshop on automated vehicles and law enforcement. Invited experts were given specific scenarios to inform discussion of four types of law enforcement interactions:

  1. Traffic stops
  2. Collisions
  3. Emergencies (such as evacuations and detours)
  4. Tangential interactions (such as use of autonomous vehicles as evidence sources in investigations or exclusion of autonomous vehicles from zones where traffic is prohibited)

Convening the expert panel was one component of NIJ’s Priority Criminal Justice Needs Initiative. Invitations to participate reflected reviews of the literature and consultations with federal partners and law enforcement officials who have engaged on autonomous vehicle issues. The workshop focused on autonomous vehicle scenarios that had already happened on the road or could be expected to arise within five years of the July 2019 workshop.

To rank needs identified by the participants, the project team used a formula that blended weighting of the relative importance of each challenge and the probability of achieving an associated solution.

Priority Needs Identified

In all, the panel identified 33 priority needs with 17 deemed “top-tier needs,” falling under the categories Cybersecurity and Autonomous Vehicle Communication, Stakeholder Communication and Collaboration, and Standard Procedures, Guidelines, and Training. (See lists of 33 priority needs and 17 top-tier needs in Tables A.3 and 1, at pages 21 and 9, respectively, of report “Autonomous Vehicles and Law Enforcement,” linked below in the “About This Article” section.)

Top cybersecurity and autonomous vehicle communication needs included (in addition to developing ways for law enforcement to know when a vehicle is operating autonomously): developing ways to check autonomous vehicle documentation; facilitating communication with a responsible human during an autonomous vehicle traffic stop or emergency; and developing ways to securely direct autonomous vehicles or communicate intentions to them. The panel noted that electronic signaling capability could introduce a cybersecurity risk.

Key stakeholder communication and collaboration needs included: educating law enforcement and the autonomous vehicle industry on their respective processes, requirements, and technical capabilities; developing access to vehicle information that could benefit law enforcement, such as data collected by a vehicle that could aid in an incident response or investigation; and developing technology to facilitate emergency vehicle access.

Priority needs for standards and training identified by the panel included developing guidelines for dealing with inoperative autonomous vehicles, as well as guidelines on autonomous vehicle behavior during traffic stops by officers.

Conclusion

The advent of autonomous vehicles filling the roads will take longer than first anticipated, but the societal change will be massive. The expert panel on autonomous vehicles and law enforcement agreed that as the technology rapidly advances in coming years, law enforcement agencies must accelerate their preparation for a radically altered reality on the roads.

About this Article

The research described in this article was funded by NIJ award 2018-75-CX-K006, awarded to the RAND Corporation as part of the Priority Criminal Justice Needs Initiative (PCJNI). PCJNI is a project of RAND, RTI International, the Police Executive Research Forum, and the University of Denver, under direction of the National Institute of Justice. This article is based on the grantee report Autonomous Road Vehicles and Law Enforcement (2020), by Sean Goodison et al.

Date Published: February 16, 2021