CDT provided recommendations to the Departments of Justice and Homeland Security on transparency and policy around federal law enforcement use of facial recognition, in response to an October 13 request by the agencies for written submissions.
In our comments, we describe a set of recommended best practices divided into five categories: procurement, transparency, deployment, redress, and accountability. We urge the Department of Justice (DOJ) and Department of Homeland Security (DHS) to promptly adopt these policies. While states and cities have begun to act to prevent harms inflicted by facial recognition, with over a dozen enacting meaningful limits and some jurisdictions banning the technology entirely, the federal government has unfortunately fallen behind. Neither Congress nor federal agencies have enacted regulations or policies that are critical to protect civil rights and civil liberties.
Facial recognition is a powerful surveillance technology, and absent strong rules and safeguards, it can significantly harm individuals. The risk is acute, because facial recognition is a double-edged sword: It is dangerous when it works poorly, and can be dangerous in an entirely different way when it works well. There are numerous documented cases of facial recognition causing Americans to be improperly arrested and jailed, as well as several recorded instances of facial recognition being abused to identify and catalog individuals engaged in First Amendment-protected activities. Because of the range of risks facial recognition presents, law enforcement use requires a comprehensive set of policies and limits.