Meta Fires a Warning Shot Against Police Surveillance, But It Must Do More

Rachel Levinson-Waldman is deputy director and Mary Pat Dwyer is a fellow at the Liberty and National Security Program at the Brennan Center for Justice at NYU School of Law.

Law enforcement agencies at the federal, state, and local level increasingly monitor social media. While there is little evidence that online surveillance prevents crime, there is ample evidence of its dangers. Police have unjustly monitored Black Lives Matter activists, and one police department’s social media activity helped spread a ludicrous online conspiracy theory. Social media monitoring by law enforcement chills protected speech and association, and has led to wrongful arrest.

The social media companies themselves have largely stayed out of the debate about how and to what extent police should monitor social media. So it’s significant that Meta, Facebook’s parent company, recently wrote to the Los Angeles Police Department demanding that its officers cease the use of Facebook for surveillance and signaling stronger measures in the future. Meta issued its letter in response to reporting by the Brennan Center for Justice (where both authors work) and The Guardian about how the LAPD monitors and collects information from social media.

Internal police documents obtained by the Brennan Center reveal that the LAPD permits and even encourages the use of fake accounts on the platform. The LAPD’s Social Media User Guide, for instance, allows officers to create a “Fictitious Online Persona” to study “trends,” develop profiles, and conduct “research.” No documentation or supervisor approval is necessary. An undated training video produced by the LAPD further instructs officers to “make sure [they] set up a dummy account” when looking for information on Facebook – a directive that squarely violates Facebook’s policy against inauthentic behavior on the platform.

The LAPD also tested a tool offered by Voyager Labs, a private company that collects and analyzes vast quantities of social media data, including information gathered using fake accounts. Voyager boasts that it can divine people’s motives and identify those most “engaged in their hearts” with their own ideologies. For instance, in a case study provided to the LAPD and disclosed to the Brennan Center, Voyager touts its ability to capture biographical information from a target’s Facebook friends, as well as friends of friends with no direct connection to the original subject, and collect it in a searchable dataset that police could use for further monitoring. Voyagers’s analytics tools clearly violate Facebook’s "Platform Terms," which bar developers from using Facebook data to “perform, facilitate, or provide tools for surveillance,” including for law enforcement or national security purposes.

Meta’s letter to the LAPD affirmed that the use of fake accounts violates its terms of service. The company also demanded more generally that the LAPD cease collection of data for surveillance purposes. These pronouncements serve notice to police departments nationwide that they are not exempt from Facebook’s policies. In an apparent response to the Voyager allegations, the letter also reiterated Meta’s prohibition on developers’ scraping data from the platform for surveillance purposes. Finally, Meta promised to take further steps to counter surveillance, saying it will disable fake accounts and take action against vendors that violate the terms of service.

This letter is an important public statement, particularly given the growing number of companies, police departments, and federal law enforcement agencies that use social media to monitor protests and demonstrations without any reason to suspect criminal activity. It also sets an example for smaller companies that currently collect vast amounts of individual information while providing less transparency than Meta does into their standards of conduct and enforcement decisions.

While Meta’s letter is an important first step, it must not be the company’s last. To date, the company has been more aggressive in acting against academic researchers studying misinformation with users’ consent than with developers collaborating with law enforcement agencies. The Brennan Center’s list of known social media monitoring products would be a good starting point to identify developers that exploit user data for surveillance purposes. Meta should also disable fake accounts created by the LAPD using the tools it uses to identify and remove other accounts associated with “coordinated inauthentic behavior.”

Meta could go even further by limiting or banning the LAPD’s use of Facebook, much as the platform has done with other violators, to the extent possible without unduly hampering Angelenos’ ability to communicate with the police department.

Other social media companies should likewise take steps to identify and remove these abusers from their platforms. Twitter, for example, prohibits the use of its platform for surveillance purposes and could remove vendors on that basis. To date, however, the company’s responses to surveillance companies’ use of its data have been mixed. Today it announced it had suspended the developer account of ABTShield in response to inquiries by the Guardian on its collection of data for the LAPD. However, Twitter has not responded to the Voyager allegations; and in 2020, the company defended its decision not to remove Dataminr from its platform – a decision it apparently has not revisited despite ongoing reporting that law enforcement agencies, including LAPD, have used Dataminr’s tools for surveillance.

Meta’s letter is a powerful tool for activists and civil society groups seeking to limit the use of social media for surveillance. It sends a clear message about the unacceptability of fake accounts and the use of the company’s platforms for surveillance and monitoring – both of which appear to be widespread practices, though the dearth of robust, publicly available police department policies makes a full accounting impossible. The passage of transparency and accountability legislation, such as the ACLU’s Community Control Over Police Surveillance model bill, would help compel the production and disclosure of such policies, particularly when coupled with robust oversight mechanisms.

We are gratified that our transparency efforts – and our litigation battle with the LAPD – helped elicit Meta’s important letter. To be more than an empty threat, however, it must be followed by action. The company should enforce its policies against both law enforcement agencies and developers, and other social media platforms should follow suit.