Much of the planning for the Jan. 6 attack on the U.S. Capitol Building took place in the open, on mainstream and far-right social media platforms available to anyone, raising the question of why the law enforcement response at the Capitol was so woefully insufficient. The event itself erupted in plain sight as well, with the perpetrators simultaneously plastering their social media with pictures, videos, and brags about their participation.
In response, the FBI has undertaken a wide-ranging investigation, relying on these social media posts to help identify and track down suspects. With the potential for additional armed confrontations, we are likely to see calls to further ramp up law enforcement’s use of social media to identify violent actors and prevent further attacks.
While social media analysis will be critical to these investigations, dragnet social media monitoring brings significant risks, and even targeted monitoring can be abused. As with other forms of surveillance, social media monitoring has been used in opaque and abusive ways, and it is likely to disproportionately target activists and communities of color without stringent limitations in place. Social media can also be highly contextual and susceptible to misinterpretation. While many posts leading up to the events of Jan. 6 revealed an explicit intent to engage in violent, unlawful, and profoundly antidemocratic activity, including carrying weapons across state lines to threaten and even kill elected officials, innocuous online sentiments can also be perceived as a threat, with grave consequences for the poster. Government officials will therefore need to tread carefully when looking to social media to predict bad acts.
With threats on social media likely to increase, as well as the real risks of misuse of social media to target activists and other individuals, how far can federal, state, and local law enforcement agencies go in surveilling social media, and what is the appropriate scope of law enforcement access?
FBI policies
There are no constraints – in law or internal rules – on FBI agents’ ability to review publicly available social media content, and agents can access or receive information from closed groups under some circumstances as well. Once criminal activity is suspected, agents can use undercover online identities, as long as they follow FBI policy regarding investigations of political organizations where necessary.
The rules for FBI investigations are set out in the Bureau’s Domestic Investigations and Operations Guide, or DIOG, which implements the Attorney General’s Guidelines for Domestic FBI Operations. These rules limit the collection of First Amendment-protected information; the Bureau is forbidden from basing its conduct “solely on an individual’s legal exercise of his or her First Amendment rights,” while permitting agents to gather information online about constitutionally protected rights if the collection is “logically related to an authorized investigative purpose,” does not interfere with the group’s constitutional exercise of their First Amendment rights, and is the least intrusive alternative that is reasonable in the circumstances.
In practice, the FBI has a long history of targeting First Amendment-protected speech, especially by Black activists and Muslim communities. In 2017, for instance, the Bureau invented a new category of threat it called “Black Identity Extremism” (BIE), which appeared designed essentially to target activists of color; the FBI eventually withdrew it under fire, while subsuming BIE under a different category that simply obscured its surveillance of Black activists. The FBI has also put entire predominantly Muslim neighborhoods under surveillance with no suspicion of criminal activity, and has conducted intensive spying operations against indigenous and other environmental activists.
Before formally opening any kind of inquiry, agents can search and review publicly available social media and other online information (see Appendix L of the DIOG). After initiating the first level of formal FBI inquiry, called an assessment, FBI agents can record and monitor “public, real-time communications,” such as conversations in a public chat room, if recording is both necessary to the assessment and the least intrusive method to obtain the information. Assessments are notoriously lax, requiring only an “authorized purpose” and a “clearly defined objective,” with no reasonable indication, or even allegation, of wrongdoing required.
During an assessment, an agent may also access a private or restricted-access social media forum, such as a private Facebook group or subreddit, if she is given access by a “consenting party” who has authority to access and control content on the site. That person could be a system administrator, or a company official who can direct others regarding the content on the site. It can also be an account holder for the site, including an administrator for a private group. Because any member of a private group could share information with an FBI agent, however, the standard does not impose robust limitations in practice. Moreover, the carve-out for company officials suggests that the FBI could obtain consent from, for instance, certain high-level Twitter, Facebook, or Reddit officials or moderators. An agent cannot, at the preassessment or assessment stage, use a fictitious identity to gain access or connect with individuals online.
Once there is a reasonable indication of criminal activity – which would certainly include an armed attack on the Capitol – an agent may monitor real-time communications in a private chat room or another online forum without the knowledge of the forum members; the person or body who must authorize the monitoring is redacted from the DIOG. The agent must make a two-part showing to do so: first, that the information to be obtained is relevant to the investigation, and second, that “recording the communications is the least intrusive method to obtain the evidentiary information, weighing the investigative value of the evidence to be obtained against the potential collection of First Amendment information.”
An agent may also create an undercover social media account – for instance, to pose as an enthusiast seeking to become involved in the cause – after receiving approval from a Special Agent in Charge (or, in some instances, the FBI headquarters). Once undercover, the FBI employee must take all necessary and reasonable steps to protect potential victims and prevent serious criminal activity if online contact reveals a public safety threat.
When the undercover activity involves a political organization, the agent must solicit approval from FBI headquarters, with an application that includes the length of the operation and “a statement of which circumstances are reasonably expected to occur, what the facts are likely to be, and why the undercover operation merits approval in light of the circumstances.” The Guidelines do not define the term “political organization,” and it is unclear whether the term would cover the Proud Boys. The group, whose leaders have celebrated violence as “a really effective way to solve problems,” has been at the center of previous clashes and earned itself a spot on the Southern Poverty Law Center’s list of hate groups. At the same time, it identifies itself as a body dedicated not to criminal acts but to “reinstating a spirit of Western chauvinism,” and its members have strongly aligned themselves with President Donald Trump, who famously told them during a presidential debate to “stand back and stand by.” Reports that the FBI had designated the Proud Boys as an extremist group in late 2018 were subsequently refuted.
State and local law enforcement policies
State and local law enforcement agencies have also established their own social media monitoring programs. A 2016 survey of 539 police departments by the International Association of Chiefs of Police and the Urban Institute found that 70 percent of police departments use social media for gathering intelligence and monitoring public sentiment. Another 59 percent of departments reported having contacted Facebook or Twitter to obtain information to use as evidence. Like the FBI, some police departments also allow the creation of undercover accounts to conduct surveillance, including those in Austin, Los Angeles, New York, and Salt Lake City.
Some big city police departments, including the NYPD and LAPD, have policies governing online undercover activity, though the NYPD has been heavily critiqued for its online monitoring. The vast majority of departments engaging in social media monitoring do not have publicly available social media policies, and very few policies recognize and account for the First Amendment rights of individuals under scrutiny. Law enforcement agencies have also been known to contract privately with social media analytics firms like Dataminr, which analyze public data feeds and deliver automated alerts to law enforcement clients, including the FBI, about protests, shootings, and natural disasters.
Platform policies
The major platforms, including Facebook, Twitter, and Reddit, typically require a subpoena or court order for basic subscriber information (such as name, email addresses, and IP addresses of recent logins) and a search warrant for contents of communications, including videos, photos, and direct messages. They also set out a process for law enforcement officers to follow in emergency situations, such as where there is danger of death or serious injury. While law enforcement policies for Parler, the darling of the right wing (at least up until recently), are not available online, the social network told Amazon it had been “actively working with law enforcement for weeks” before the Internet giant took it down.
Facebook, Instagram, and Twitter also have policies in place prohibiting use of their user data for surveillance, while Facebook has a “real name” policy that demands that users identify themselves online as they would in real life, with no exception for law enforcement activity. After a 2018 lawsuit exposed the Memphis Police Department’s use of fake profiles to collect Black Lives Matter activists’ private posts, Facebook sent a letter to the agency requesting that they “cease all activities on Facebook that involve impersonation.”
Constitutional limitations on social media monitoring
Courts have typically held that individuals do not have a reasonable expectation of privacy in data publicly shared online, with one appeals court likening posting a tweet to “scream[ing] it out the window.” This understanding stems from the third-party doctrine, which generally holds that individuals do not have an expectation of privacy – and thus do not have Fourth Amendment rights – in information shared with a third party, even if they did not intend the data to be shared with the government. Thus, one district court concluded in 2012 that when a Facebook user posted messages on his profile, he no longer had an expectation of privacy in that data, even if he had enabled privacy settings limiting access to his posts. While he “undoubtedly believed that his Facebook profile would not be shared with law enforcement,” his “legitimate expectation of privacy ended when he disseminated posts to his ‘friends’ because those ‘friends’ were free to use the information however they wanted — including sharing it with the Government.”
Courts could, however, be amenable to limiting social media monitoring where it implicates First and Fourteenth Amendment freedoms. While the Supreme Court has held that a governmental surveillance program can’t be challenged solely on the grounds that it creates a chilling effect, at least one federal appeals court has ruled that where surveillance is intentionally directed at a protected class, it may give rise to a colorable claim. If social media is used to intentionally target minority communities in ways that infringe upon their First Amendment-protected rights, or is used to retaliate against people or groups for exercising their First Amendment rights, they may be able to challenge it successfully.
Next steps
As this piece sets out, the FBI asserts broad authority to use social media both to investigate past crimes and to prepare for potential coming unrest. State and local law enforcement agencies have claimed wide authorities as well. Thus far, with the exception of one consent decree arising from police department misconduct, no court has imposed limits on law enforcement’s use of social media monitoring, and law enforcement agencies’ broad policies impose few constraints. At the same time, as with other forms of surveillance, social media monitoring has been used in opaque and abusive ways, and it is likely to disproportionately target activists and communities of color without stringent limitations in place.
Policymakers should therefore consider instituting the following measures:
- Any law enforcement agency using social media for surveillance or criminal investigations should have a policy in place that is easily available to the public. Jurisdictions should establish a democratic mechanism to invite and incorporate community members’ and experts’ input into the content of the policy.
- State and local law enforcement agencies should analyze and disclose the scope and impact of their social media data collection, including the types of investigations in which the practice is used and its effect on protected constitutional rights, as well as the existence of measures to protect the public from the adverse effects of social media monitoring.
- Agencies should prohibit the collection of social media data based on an individual’s or group’s exercise of First Amendment rights; their race, religion, ethnicity, or other category protected by law; or actual or perceived immigration status.
- Where social media is assessed to prepare for a pending event, its use should be carefully limited to ensure that a public safety justification is not transformed into license for broad-scale monitoring. Monitoring in advance of an event – separate from social media analysis related to a criminal investigation, described below – should be undertaken only to make determinations about the resources necessary to keep participants and the public safe, and only where there are articulable and credible facts describing the public safety concerns justifying the monitoring. There should be a process in place to document those facts in writing and to obtain supervisory approval, with records maintained solely to facilitate audits and other accountability measures.
- When social media monitoring is used as part of a criminal investigation, it should be on the basis of specific and articulable facts showing reasonable grounds to believe that the data sought is relevant and material to an ongoing criminal investigation. Information collected should not be retained unless it is determined by a supervisor to be relevant to an ongoing criminal investigation.
- Finally, while there are currently no constitutional limits and few policy limitations on the use of undercover social media accounts, undercover online activity is sufficiently susceptible to abuse that the use of such accounts should require a warrant or similar protections. The Title III process for wiretap warrants offers a model that could be adapted for social media warrants, including a requirement that no less intrusive alternative be available.
These measures would create much-needed guardrails around law enforcement access to social media, while also preserving legitimate law enforcement preparation and investigation of criminal activity.