Data Mining License Plate Records
On Friday I posted about an ongoing effort by the DEA to put automatic license plate reading (ALPR) devices on public interstates, where they will sweep up records of Americans’ travel and store it for two years. The agency is now pushing to deploy them in Utah and has already done so in states along the southern U.S. border.
But looking forward, if nothing is done, we can expect this technology to expand not only geographically, but also in how it is used. In particular, we can expect all kinds of cross-referencing and data mining techniques to be applied to the information streams generated by ALPR scanners.
As I noted in Friday’s blog, the DEA has said that it wishes to use the technology for “intelligence” and to “research the movements” of suspects, and for “statistical information.” All this points toward efforts to data mine license data.
We know that ALPR is already being used this way in at least one government program: the Automated Targeting System, which assigns computer-generated “risk assessment” scores to all travelers who cross the nation’s borders, and retains the data and scores for 40 years (we wrote up these materials on ATS in 2007 when the program first came into public view). According to DHS, in a 2006 ATS Privacy Impact Assessment,
ATS-Land (ATS-L) is a module of ATS that provides for the analysis and rule-based risk assessment of private passenger vehicles crossing the nation's borders. By processing and checking of the license plate numbers of vehicles seeking to cross the border, ATS-L allows CBP officers to cross-reference the TECS [a law enforcement database] crossing data, TECS seizure data, and State Department of Motor Vehicle (DMV) data to employ the weighted rules-based assessment system of ATS. In this way ATS-L provides, within seconds, a risk assessment for each vehicle that assists CBP Officers at primary booths in determining whether to allow a vehicle to cross without further inspection or to send the vehicle for secondary evaluation.
Local law enforcement will undoubtedly want to engage in similar “risk analysis” cross-referencing—especially as ALPR, combined with other technologies, makes it increasingly easy to do so. If the technology continues to get cheaper, and no restrictions are placed upon police use of it, and it ends up deployed on every block, we could easily see scenarios such as:
Over time, as stories of such stops spread, Americans begin to feel that they are being watched wherever they drive and must constantly ask themselves, “Am I doing anything suspicious?” When they watch movies of the old days when people could drive around anonymously, they feel nostalgic for that lost freedom.