On the morning of December 4 last year, precisely at 8:22 am, a vehicle driving along a quiet residential street in Alabama activated its license-plate-reading cameras, capturing images of the cars it passed. One notable image, devoid of any vehicles or license plates, features a vibrant red “Trump” campaign sign prominently displayed in front of a garage. In the backdrop, a banner mentioning Israel can be seen alongside a holly wreath and a cheerful inflatable snowman.
On another occasion, a different vehicle captured another image displaying a “Steelworkers for Harris-Walz” sign planted in a front yard. This image reveals a construction worker, clearly visible, positioned near an additional Harris sign. Further photos depict a range of bumper stickers, including those supporting Trump and Biden (and even a “Fuck Biden” sticker) affixed to various trucks and cars across the nation. One snapshot, taken in November 2023, shows a bumper sticker advocating for the Obama-Biden ticket that is partially torn.
These visuals originated from AI-enhanced cameras installed in cars and trucks, which were originally intended for reading license plates. However, they have now expanded to document political signs in front of private residences, individuals wearing themed T-shirts, and vehicles adorned with pro-abortion bumper stickers—all while meticulously recording the exact locations of these sightings. Recently acquired data reviewed by WIRED highlights how a tool that was originally meant for traffic oversight has transformed into a mechanism capable of monitoring free speech protected by the US Constitution.
The collection of detailed photographs emerged in results generated by the services of DRN Data, a company specializing in license-plate recognition (LPR) and owned by Motorola Solutions. This LPR system is utilized by private investigators, repossession agents, and insurance firms. Additionally, a related Motorola division, known as Vigilant, enables law enforcement to access the same LPR data.
Files provided to WIRED by artist Julia Weist, who is documenting restricted datasets as part of her project, reveal how individuals with access to the LPR system can search for specific phrases or names, such as those of public figures, and receive photographs where the search term appears, regardless of whether it is visible on the license plates.
A query using the license plates of Delaware vehicles for the term “Trump” returned over 150 images depicting private residences and bumper stickers. Each result also includes the date, time, and precise location of the captured photograph.
“I searched for the word ‘believe,’ and all it returned were lawn signs. There were messages painted on planters along the roadside and even someone wearing a sweatshirt that says ‘Believe,’” Weist explains. “When I searched for the word ‘lost,’ it pulled up flyers that people post for missing dogs and cats.”
This research not only highlights the extensive capabilities of LPR technology, which has accumulated billions of images of license plates, but it also underscores how individuals’ political opinions and their residences can be stored in massive databases that are accessible for searching.
“It truly highlights how extensive surveillance is occurring on a large scale in the peaceful streets of America,” remarks Jay Stanley, a senior policy analyst at the American Civil Liberties Union. “This surveillance extends beyond merely license plates; it encompasses a wealth of other highly revealing information about individuals.”
In a statement to WIRED, DRN asserted that it adheres to “all applicable laws and regulations.”
License plate recognition systems operate by first capturing an image of a vehicle, then employing optical character recognition (OCR) technology to pinpoint and extract the text from the license plate within the captured image. DRN, owned by Motorola, offers a variety of license-plate-recognition cameras: a stationary camera that can be placed alongside roads to identify a vehicle’s make and model and capture images of cars traveling at speeds up to 150 mph; a “quick deploy” camera that can be mounted on buildings to keep track of vehicles at various properties; and mobile cameras that can be fitted on dashboards or mounted to vehicles to capture images while being driven.
Over the course of more than ten years, DRN has compiled over 15 billion “vehicle sightings” throughout the United States. The company claims in its promotional content that it gathers over 250 million sightings each month. Images in DRN’s commercial database are made available to law enforcement through its Vigilant system, while photographs collected by law enforcement themselves are not integrated back into the larger database.
The system relies in part on DRN “affiliates” who set up cameras in their vehicles, including repossession trucks, capturing license plates as they travel. Each vehicle can accommodate up to four cameras, recording images from various angles. These affiliates earn monthly bonuses and can also obtain free cameras and search credits.
In 2022, Weist achieved certification as a private investigator in New York State. This certification granted her access to a vast array of surveillance software available to PIs, including the DRN analytics platform, DRNsights, as part of her affiliation with investigations company IRBsearch. (However, after Weist published an op-ed detailing her work, IRBsearch audited her account and subsequently discontinued it. The company did not respond to WIRED’s request for comment.)
“There is a difference between tools that are publicly accessible, like Google Street View, and things that are searchable,” Weist remarks. During her investigations, she performed multiple searches using various terms, yielding results that exceeded mere license plates. She shared data with WIRED showing that a search for “Planned Parenthood,” for instance, revealed stickers on vehicles, on bumpers, and in windows, expressing a range of sentiments towards the reproductive health services organization. Civil liberties advocates have already voiced concerns about how license-plate-reader data could potentially be weaponized against those seeking abortion.
Weist expresses her worries over how these search tools could be misused amid rising political violence and societal division. Although not directly related to license plate data, an Ohio law enforcement official recently stated that individuals should “write down” the addresses of those who display yard signs supporting Vice President Kamala Harris, the Democratic presidential nominee for 2024. This exemplifies the potential for a searchable database of citizens’ political affiliations to be misused.
A 2016 report by the Associated Press revealed widespread misuse of confidential law enforcement databases by police officers across the country. In 2022, WIRED uncovered that hundreds of US Immigration and Customs Enforcement employees and contractors faced investigations for abusing similar databases, which included license plate recognition (LPR) systems. The alleged misconduct highlighted in both reports ranged from stalking and harassment to the sharing of information with criminals.
While individuals use lawn signs or bumper stickers to express their views and influence those around them, the ACLU’s Stanley emphasizes that this expression is meant for “human-scale visibility,” rather than for machine tracking. “They may wish to convey their beliefs within their communities and to their neighbors, but they don’t want to be recorded in a nationwide database accessible by law enforcement,” Stanley remarks.
Weist argues that the system should at least be able to filter out images that do not include license plate data, avoiding errors. “Any number of mistakes is unacceptable, particularly when it involves trivial details like people’s clothing or lawn signage,” Weist asserts.
“License plate recognition (LPR) technology aids public safety and community services, assisting in locating abducted children and stolen vehicles, as well as automating toll collection and reducing insurance costs by combating fraud,” states Jeremiah Wheeler, the president of DRN.
Weist suggests that, considering the limited number of images featuring bumper stickers relative to the countless vehicles displaying them, Motorola Solutions might be trying to filter out images that include bumper stickers or similar text.
Wheeler did not provide answers to WIRED’s inquiries regarding the potential limitations on searches within license plate databases, the presence of images depicting homes with lawn signs but lacking visible vehicles in search results, or whether any filters are implemented to minimize such images.
“DRNsights adheres to all relevant laws and regulations,” Wheeler states. “The DRNsights tool provides authorized individuals access to license plate data and related vehicle details that are collected in public areas and observable by all. Access is limited to clients with specific legitimate purposes in accordance with the law, and those who violate this agreement will have their access revoked.”
In recent years, license plate recognition systems have seen significant growth due to advancements in camera technology and improvements in machine learning techniques. These systems, including DRN and its competitor Flock, signify a shift in the manner in which individuals are monitored as they navigate through urban and suburban environments.
In recent times, CCTV cameras are increasingly being outfitted with artificial intelligence to observe individuals’ movements and even interpret their emotions. These systems hold the promise of alerting officials who may not be able to continuously oversee the CCTV footage, responding swiftly to real-world situations. However, the efficacy of license plate recognition in crime reduction remains a subject of skepticism.
“When government or private enterprises promote license plate readers, they often present it as if the technology is exclusively seeking lawbreakers or individuals suspected of car theft or involved in amber alerts, but that is not reflective of how the technology operates,” states Dave Maass, the director of investigations at the civil liberties group Electronic Frontier Foundation. “This technology accumulates data from everyone and retains that information for often extensive periods.”
As time progresses, the capabilities of such technology may expand. Maass, who has extensively researched license plate recognition systems, notes that companies are now exploring “vehicle fingerprinting,” a method that identifies the make, model, and year of a vehicle based on its shape and can even assess if the vehicle has damages. According to DRN’s product pages, one forthcoming update will enable insurance companies to verify whether a vehicle is being utilized for ride-sharing purposes.
“The framework of the country is designed to protect citizens from government encroachment, yet there are minimal safeguards to protect us from private entities engaged in profit-driven ventures,” remarked Nicole McConlogue, an associate professor of law at the Mitchell Hamline School of Law, who has investigated license plate surveillance systems and their potential for bias.
“The extent to which they are able to do this is particularly concerning,” McConlogue states regarding vehicles traversing streets to capture images. “By doing so, you also embody the motivations of those who are gathering the data. Moreover, in the United States, it also carries the historical consequences of segregation and redlining, which have significantly impacted the makeup of neighborhoods.”