Where Is the Surveillance Fight Headed?
Towards stricter legislation in the United States in the wake of broader use of facial recognition technologies
“Quis custodiet ipsos custodes” (“Who will watch the watchmen?”). The phrase by Roman poet Juvenal was ominously sprayed in graffiti across the walls of cities that filled with protesters during the most recent George Floyd protests.
This past summer’s protests for police reform have engendered increasingly strict calls for restricting the use of facial recognition technology by police, a turn away from escalating surveillance measures under the Trump administration.
In a compound republic, the fight over surveillance technology laws is a fragmented affair, with varying restrictions and freedoms across the states. Few states have imposed restrictions on governmental use and virtually no states have set up guardrails against commercial use of the technology. Illinois was the first state to introduce laws restricting use of the technology in accordance with its “Biometrics Privacy Act,” which arose when fingerprinting technology became popularized in 2008 and concerns were rampant over identity theft Texas followed suit in 2009 with prohibitions on selling biometric information without users’ consent. And then there was a long interlude until the biggest state in the union and the home of Silicon Valley introduced the California Consumer Privacy Act (CCPA), akin to the European Union’s own GDPR. At first, many states did not follow suit in California’s wake; this trend drastically changed when the Covid-19 crisis hit and protests this year spawned civil society groups to push for more restrictions on facial recognition technology.
Today’s legislative battle has devolved into a tug-of-war between what I refer to as ‘strict’ and ‘loose’ legislation. It is a common misconception embedded in the rhetoric about facial recognition technology that states have only two options: either to ban the technology to not ban it. Arguments put forth on either side echo the claims that we should ban the technology because of its inherent biases, or we should not ban it because of its benefits for security and law enforcement. Opponents voice concern with the misuse of the technology by law enforcement to surveil and police peaceful protesters, restricting their First Amendment rights, and the higher false positive error rate of the technology as it relates to incorrectly flagging minorities. Whereas proponents are quick to insist that the technology is more accurate than manual, human processes, saving law enforcement valuable time and money.
Yet, legislative efforts to regulate the use of facial recognition technology do not reflect the diametrically opposed sides of proponents and opponents. Rather, there is great variance in recent efforts to regulate the use and distribution of the technology. To better reflect this variance in legislative proposals, I suggest a scalar reading of ‘loose’ to ‘strict’ legislation, in contrast to the binary approach of ‘ban’ or ‘no ban’. Looseness can be broadly defined as regulations that include ‘more exceptions’ — this loose legislation notably contains many “unless” clauses, such as “businesses may not collect and store biometric information unless they receive notice and consent for ‘security purposes’. An extreme level of looseness means a wholesale security exception has been built into the legislation, as is the case with recently proposed laws in Washington and Michigan.
Strictness, on the other hand, can be broadly defined as regulations with more stringent controls — this strict legislation notably contains many “under no circumstances” clauses, such as “under no circumstances may the municipality obtain, retain, access or use any face surveillance technology”. An extreme level strictness means there is an outright ban or moratorium imposed on law enforcement’s use of the technology, as is the case with Boston’s or Oakland’s municipal bans.
It is striking to note that ‘strict’ legislative proposals so far have all passed in state legislatures (as in New Hampshire and Washington), whereas the ‘loose’ legislative proposals (as in South Dakota and Utah) have all so far failed. Congress has matched this trend among state legislatures to propose stricter legislation, with Senators Markey and Merkley along with Representatives Jayapal and Pressley sponsoring the most recent Moratorium Act in late June of this year.
I hypothesize that this trend will continue and that we will see stricter legislation proposed and passed in the coming year, both at the federal and state level. The question remains, what will be the societal repercussions of this legislative trend, and should we welcome it?
Thank you to the faculty at the Berkman Klein Center for Internet and Society at Harvard University for valuable discussions on this research at the Summer Institute.