Can Flock Cameras Ever Be Trusted? UVA Cybersecurity Expert Weighs In on Privacy Safeguards for License Plate Readers
University of Virginia School of Data Science's Daniel Graham discusses warrant requirements, data retention, data deletion, and audits as ways to build public confidence in license plate reader technology
Use of automatic license plate readers is growing across the country, and so are concerns about mass surveillance. The Charlottesville City Council discontinued its contract with Flock Safety after a one-year pilot of 10 cameras. Proponents cite law enforcement benefits, while critics argue the technology enables broad surveillance without adequate privacy protections.
Daniel Graham, associate professor at the University of Virginia School of Data Science and author of the bestselling cybersecurity book Ethical Hacking, conducts research at the intersection of cybersecurity, software engineering, and data science. In a Q&A with the School, he discusses why license plate readers have sparked public concern and what technology companies could do to better protect privacy.
Graham says the cameras' connectedness is what makes them a challenge. Because they share information across locations, they can be used to establish trends in people's movements. The data is stored in a single location, and anyone with access to it can query the cameras without a warrant. Each camera processes images on the device and securely sends the results to a central database.
Graham compares this to cell phone location data. He notes that in Carpenter v. United States, the Supreme Court determined that access to historical cell-site location records should generally require a warrant, and that software gates have been put into those systems to prevent warrantless access. The same software safeguards are not currently in place for Flock cameras, he says.
Graham suggests building public confidence through:
A warrant and due process requirement for accessing camera data, as with cell phone location records
Shorter retention. He floated five days instead of the current 30 for data not useful to a case.
The ability to delete one's data, similar to controls over Google search history
Audits by ethical hackers and cybersecurity firms, comparable to financial audits
On reports of misuse, Graham says it can occur when built-in protections aren't turned on, which can give users broader access than intended, such as repeated searches of the same plate or access to cameras outside their region.
He describes three levels of safeguards: users who deploy the systems configuring them correctly, vendors designing systems that make misconfiguration harder, and policy that keeps pace with reasonable expectations of privacy.
"I think ethical hackers can help from the technical perspective, but we really need policy thinkers to shape these ideas that we want to implement," Graham says.
