AVs Must Steer Clear of Illegal Surveillance

In light of the 2021 Law and Mobility Conference’s focus on equity, the Journal of Law & Mobility Blog will publish a series of blog posts surveying the civil rights issues with connected and autonomous vehicle development in the U.S. This is the fourth and final part of the AV & Civil Rights series. Part 1 focuses on Title VI of the Civil Rights Act. Part 2 focuses on the Americans with Disabilities Act. Part 3 focuses on Title II of the Civil Rights Act.

Your data says a lot about you, and widescale adoption of connected and automated vehicles (AVs) will create mountains of data that says even more. Based on location data alone, AV companies may know where you live, where you work, what you like to do in your free time, who you hang out with, and possibly even your religious and political beliefs. And, again, this is just based on location data; AV companies will also have extensive records of your biometric and financial information. Overall, AVs can provide constant and near-comprehensive surveillance. So what happens when the government gets access to surveillance collected by private companies and relies on it in criminal investigations?

This was the real-life nightmare of Robert Julian-Borchak Williams, a Black man in the Detroit area who was arrested and charged for a crime that he didn’t commit. A facial recognition algorithm employed by the Detroit Police Department matched his face to a surveillance image from a robbery. Williams is known as the first person arrested as the result of a bad algorithm.

As of fall 2020, at least 360 police departments use facial recognition technologies, 24 use automated data analysis tools, and 26 use predictive policing measures, which aim to identify crimes before they happen by relying on historical data (which has been shown to be racially discriminatory and ineffective). Over 1,000 police departments use surveillance drones, which were deployed to track down and arrest Black Lives Matter protestors last summer. If AVs become a part of this network of technologies, this surveillance will become even more invasive, particularly for Black passengers and pedestrians that both police and artificial intelligence tend to manifest bias towards.

This trend of warrantless surveillance is constitutionally dubious. The Fourth Amendment protects U.S. persons from “unreasonable searches and seizures” without a warrant. Courts have considered surveillance to be an “unreasonable search or seizure” if it invades a reasonable expectation of privacy. The scope of Fourth Amendment protections was narrowed in United States v. Miller and Smith v. Maryland, where the Supreme Court held that there is no reasonable expectation of privacy if information is purposefully provided to third parties. In these cases, the Court held that the government could obtain bank records and transactional phone call data without a warrant because that information was consensually relayed to third parties (namely, bank and phone companies). However, this third party doctrine was abrogated in 2015 in United States v. Carpenter, when the Supreme Court held that a warrant was required to collect over four and half months’ worth of cell-site location information (CSLI) for the defendant, a robbery suspect. The Court noted that, if third party doctrine were applied to CSLI, “[o]nly the few without cell phones could escape . . . tireless and absolute surveillance.”

The Carpenter framework could be applied to AVs, based on the potential comprehensiveness of AV surveillance; the intimate information that AV surveillance could reveal; how cheap it would be for the government to rely on AV for both ongoing and retrospective surveillance; and the questionable level of voluntariness through which AV users would “provide” their information to companies. Thus, the application of the Carpenter framework to AV by judges is one way to avoid the incorporation of AVs into the surveillance state. On the legislative side, we could see a comprehensive federal privacy bill soon. However, policing is squarely in the state and local purview, and it is hard to say how much a federal law could reach into these surveillance issues as both a legal and a practical matter.

States like California and Virginia already have comprehensive privacy laws on the books that protect consumers with certain rights, including the right to know what data is being collected about them by large, private companies, and the right to opt out of the sale of personal information by these companies. However, government and nonprofit entities are explicitly exempt from these laws. Moving forward, privacy laws should include protection from government overreach, not just corporate overreach.

AV companies can certainly take action as well. Around the time of the Williams arrest, Amazon, Microsoft, and IBM announced that they would pause or stop offering their facial recognition data to law enforcement. These moves were largely symbolic, as police mostly rely on companies that are not household names for their data. In the AV context, company policies against warrantless surveillance and partnerships with law enforcement could provide users with some peace of mind.

Throughout my Civil Rights Series, I have emphasized the importance of data transparency so that agencies like the Department of Justice’s Civil Rights Division can easily track and investigate discriminatory impacts of AVs. “Transparency,” however, cannot be a mechanism for extending the surveillance state to these vehicles. Our increasingly connected and data-driven transportation systems cannot throw our privacy rights under the (connected and automated) bus.