A decade ago, in Packingham v. North Carolina (2016), the Supreme Court observed that social media sites are, for many, “the principal sources for knowing current events, … speaking and listening in the modern public square, and otherwise exploring the vast realms of human thought and knowledge,” affording “perhaps the most powerful mechanisms available to a private citizen to make his or her voice heard.” Despite the sensitivity of online interchanges, however, a wealth of data from social media—posts, likes, group memberships, details about organizing and political views, and identities of online connections—is easily available to government agencies who are inclined to look, as long as it is not hidden behind privacy settings. Law enforcement and homeland security agencies have long taken advantage of this easy access, using social media to surveil people based on their political beliefs or otherwise intrude upon First Amendment rights.

But the use of social media to target dissenters and opponents has soared since January 2025. President Trump and high-ranking officials have instituted an almost breathtaking set of executive orders, policies, directives, and law and immigration enforcement actions explicitly designed to evaluate, punish, detain, and deport people on the basis of their speech on social media.

To take some examples: In March, Secretary of State Marco Rubio issued a cable advising consular officers to begin scanning the social media of applicants for student or academic exchange visas to look for “conduct that bears a hostile attitude toward U.S. citizens or U.S. culture”—a measure that appeared designed to target participants in protests or scholarship advocating for Palestinian rights. The following month, U.S. Citizenship and Immigration Services (USCIS) announced that it had adopted “social media vetting for anti-Americanism,” an undefined phrase nearly guaranteed to chill expression. In June, the State Department advised applicants for academic visas that they were required to set all of their social media accounts to “public.” And evidence revealed through the Knight Institute’s legal challenge on behalf of AAUP and MESA demonstrated that DHS and the State Department explicitly sought to punish foreign students and scholars for their constitutionally-protected speech, including comments and publications posted on social media.

On the non-immigration front, reporting indicates that federal agents are recording individuals observing ICE agents and identifying them, including via social media; that information may be getting added to the federal government’s data holdings on “domestic terrorists.”

Part of the reason this weaponization of social media has been so seamless is that courts and Congress (along with local and state legislatures) have done little to stop it. There are few restrictions on access to publicly available social media or limits on enticing people to connect online through misleading means. This is in large part a function of how the Fourth Amendment caselaw has evolved—with acute consequences for First Amendment rights. In 1967, the Supreme Court ruled in Katz v. United States that the Fourth Amendment protected “people, not places.” Katz also introduced the reasonable expectation of privacy test, via Justice Harlan’s famous concurrence. Courts largely continued to hold that it was not reasonable to expect privacy from the government in public places; if you could be viewed by any member of the public, you could be freely observed and even followed by a government agent.

But as technology started to supercharge the government’s data collection and surveillance capabilities, courts began to recognize that the constitutional landscape needed to keep pace. In U.S. v. Jones (2012), multiple justices signaled that they were uncomfortable with the notion of cheap and effortless long-term tracking, and Carpenter v. U.S. took on the issue more squarely, ruling in 2018 that the acquisition of more than seven days’ worth of cell site location data required a warrant. The Carpenter Court declined to squarely overturn the third-party doctrine, which underpins its permissive approach to data collection, but it implicitly called into question the doctrine’s long-term viability.

Lower courts subsequently expanded on this approach. In 2021, the Fourth Circuit struck down the Baltimore Police Department’s persistent aerial surveillance program. More recently, a district court in Virginia acknowledged that a panoptic surveillance system that could stitch together data from a variety of sources would likely implicate constitutional privacy protections, though the court concluded that using license plate readers to identify a car’s location at discrete moments in time did not violate the Fourth Amendment because the cameras did not “reveal the whole” of citizens’ movements or “permit deductions” about them.

But courts largely have not addressed the government’s unfettered access to social media, even though an individual’s social media footprint can be far more revealing than their physical movements, facilitating deductions about or outright disclosing their political preferences, religious affiliations, sexual and gender orientation, favorite haunts, immigration status, and more. For that, we can thank the third-party doctrine, which holds that if you have voluntarily shared information with one person, even in strictest confidence, you have abandoned any reasonable expectation of privacy in it, and it may be disclosed to the government. This extends to sharing information with someone you believe to be a trusted confidant but instead turns out to be a government agent or informant.

Carpenter’s contribution was to take a more nuanced view of whether data sharing is truly “voluntary” in our digital age; the Court observed that the location data at the heart of the case was shared automatically as a function of using a cell phone, not as a voluntary choice, and that owning a cell phone was essentially a requirement of modern life. By contrast, social media is voluntary by its nature, and participation in social media is not inevitable; by and large, people choose whether to join a particular platform, whether and how often to post, what groups to join, and more. As a result, courts have had little constitutional foothold to question government entities’ surveying of publicly available social media, notwithstanding the history of misuse and overreach, or to combat the deployment of alias accounts.

What to do? First, there are compelling arguments that governmental access to publicly available social media does implicate constitutional safeguards, even under our relatively laissez faire jurisprudence. As a basic matter, if government agencies punish or retaliate against people because of constitutionally protected speech, regardless of the medium in which it appears, that clearly violates the First Amendment (as the judge in the Knight Institute’s case representing AAUP and MESA reaffirmed). In addition, it is both very simple and extremely inexpensive to gather an almost limitless volume of data about an individual or group on social media using any of the easily available commercial tools; these tools automate the process and increasingly apply artificial intelligence to analyze the results. This implicates the Fourth Amendment under Jones and its progeny, which signaled that when technological tools so expand the capacity of a law enforcement agency that monitoring, tracking, or surveilling become almost effortless, doing so violates an expectation of privacy that the Founders would have deemed reasonable.

Even when it comes to accepting connection requests from alias accounts that obscure an agent’s identity, the landscape is arguably different from the seminal 1966 case of Hoffa v. U.S. While mob boss Jimmy Hoffa didn’t know he was sharing his bribery plans with an undercover officer, he likely would have been suspicious if he noticed he was confessing to a 4’11” woman with a Canadian accent and a Disney tattoo. That’s the potential scenario online, in which the person inhabiting the social media persona could be anyone at all, sitting anywhere in the world, even pretending to be multiple people at once, tailored to different targets. It strains belief to suggest that person is an ‘invited guest,’ the principle for which Hoffa stands. Some courts have begun to recognize this in limited ways, but there could be a far more concerted effort to make this case.

Second, this area is ripe for legislation. When it comes to local and state law enforcement, the Brennan Center has proposed policies that agencies could adopt and that could form the basis for model legislation to cabin agencies’ authority. Reform at the federal level is a challenge, because of the broad authorities that federal law enforcement and homeland security agencies have and their ability to twist arms on the Hill. But in light of the visible overreach of DOJ, FBI, and DHS, and the administration’s naked targeting of its ideological opponents, there may soon be more receptivity to tackling this issue, whether through legislation squarely addressing the problem or bill language that narrows access to social media data in, for instance, immigration determinations or domestic terrorism investigations.