Skip Navigation
Analysis

White House Office Rejects DHS Proposal to Collect Social Media Data on Travel and Immigration Forms

The Biden administration should also roll back similar measures such as the State Department’s collection of social media identifiers on visa forms.

April 27, 2021

Earlier this month, the Office of Inform­a­tion and Regu­lat­ory Affairs (OIRA), the White House office that reviews federal regu­la­tions, rejec­ted the Depart­ment of Home­land Secur­ity’s proposal to collect social media iden­ti­fi­ers on travel and immig­ra­tion forms. OIRA concluded that DHS did not “adequately [demon­strate] the prac­tical util­ity of collect­ing this inform­a­tion” and noted that the Muslim ban, which ordered the proposal, had been repealed.

The proposal, which the Bren­nan Center and our allies have opposed in writ­ing, asked for author­iz­a­tion to require roughly 33 million people a year to register every social media handle they have used over the past five years on any of 20 plat­forms includ­ing Face­book, Instagram, Twit­ter, and YouTube. If approved, the meas­ure would have required a wide range of indi­vidu­als to turn over their social media handles to the federal govern­ment — includ­ing people eligible for short trips to the United States without a visa, those seek­ing asylum or refugee status, and perman­ent U.S. resid­ents seek­ing to become citizens.

Halt­ing the DHS collec­tion is a big deal, and we welcome it. But it is only a first step. The Biden admin­is­tra­tion, which is now conduct­ing a review of whether collec­tion of social media iden­ti­fi­ers “mean­ing­fully improved screen­ing and vetting,” should also end the State Depart­ment’s corres­pond­ing collec­tion from about 14 million people a year who fill out its visa applic­a­tions. Like the DHS proposal, this State Depart­ment policy was under­pinned by the Muslim ban and was justi­fied with prac­tic­ally identical support­ing docu­ment­a­tion. As OIRA’s decision signals, there is little evid­ence that social media screen­ing is an effect­ive screen­ing tool. But we do know that facil­it­at­ing drag­net surveil­lance of the modern public square harms free speech and privacy, impos­ing a dispar­ate impact on people who have tradi­tion­ally borne the brunt of govern­ment profil­ing in the name of national secur­ity.

It is unsur­pris­ing that DHS was unable to demon­strate the “prac­tical util­ity” of its proposed collec­tion. In fact, the agency’s own internal tests have ques­tioned the bene­fits of using social media to screen people coming to the United States. In a 2016 trans­ition brief prepared for the incom­ing Trump admin­is­tra­tion, DHS repor­ted that in three of the four programs it used to vet refugees, inform­a­tion from social media “did not yield clear, artic­ul­able links to national secur­ity concerns,” even when an applic­ant was flagged as a poten­tial threat through other chan­nels. (The Depart­ment did not identify any derog­at­ory inform­a­tion at all from the fourth pilot.) Among other obser­va­tions, offi­cials also poin­ted out the diffi­culty of under­stand­ing “with any level of certainty” the context and reli­ab­il­ity of what they were review­ing. They concluded that “mass social media screen­ing” was a poor use of resources, taking people away from “the more targeted enhanced vetting they are well trained and equipped to do.”

Indeed, the DHS Inspector General in 2017 reviewed the Depart­ment’s social media monit­or­ing pilot programs and expli­citly stated they could not justify scal­ing the prac­tice because DHS didn’t define criteria for success against which to meas­ure the programs. As OIRA noted, if there is any evid­ence that social media screen­ing is an effect­ive screen­ing tool, the federal govern­ment certainly hasn’t provided it.

There is, however, evid­ence that social media surveil­lance discour­ages people from freely speak­ing and asso­ci­at­ing online. We have docu­mented this chilling effect in our lawsuit chal­len­ging the State Depart­ment’s collec­tion of social media iden­ti­fi­ers on visa forms, which the Bren­nan Center filed in Decem­ber 2019 along with the Knight First Amend­ment Insti­tute and the law firm Simpson Thacher. (The lawsuit has been paused as the govern­ment reviews its screen­ing programs, though the policy remains in effect.) For example, one member of a docu­ment­ary film­maker organ­iz­a­tion we repres­ent “reviewed three years of social media activ­ity and deleted posts criti­ciz­ing the current U.S. admin­is­tra­tion” because of a fear that the posts would delay approval of their applic­a­tion. Another has “all but stopped express­ing his views and inter­act­ing with others on social media” because he under­stands the govern­ment may review and monitor his posts.

People self-censor for a number of reas­ons. They reas­on­ably fear that their speech will be misin­ter­preted, espe­cially given that commu­nic­a­tion on social media is often highly context-specific and riddled with slang and jokes. This includes commu­nic­a­tions in non-verbal form that do not have univer­sally accep­ted mean­ings (for example, whether a “retweet” on Twit­ter signals endorse­ment). Such inter­pret­ive diffi­culties are magni­fied as offi­cials try to review posts in thou­sands of languages that are under­pinned by a diverse range of customs and cultural norms, or when they rely on auto­mated tools for textual inter­pret­a­tion that have error rates of twenty to thirty percent under the best of circum­stances and perform even worse when applied to non-stand­ard English.  

Online speech is also easily misat­trib­uted. Indeed, in 2019, offi­cials turned away an incom­ing Palestinian student at Harvard (before even­tu­ally bowing to public pres­sure to let him back into the coun­try) after find­ing posts on his social media timeline that were crit­ical of the U.S. govern­ment. Notably, the posts were made not by him but by people on his friend list, and he had not inter­ac­ted with those posts. This exper­i­ence also high­lights why people may refrain from being crit­ical of the govern­ment online when they are subject to surveil­lance, since they have legit­im­ate reason to fear being penal­ized for their (or others’) speech.

Finally, social media offers a flood of sens­it­ive inform­a­tion about a person that would not be appar­ent from an immig­ra­tion bene­fit applic­a­tion and is irrel­ev­ant to what they are being screened for. While State Depart­ment policy, for example, prohib­its offi­cials from consid­er­ing certain attrib­utes to deny visas (“visas may not be denied on the basis of race, reli­gion, ethni­city, national origin, polit­ical views, gender, or sexual orient­a­tion”), offi­cials enjoy broad discre­tion to adju­dic­ate immig­ra­tion bene­fits and do so behind closed doors. This is a recipe for subject­ive bias to infect decision-making and drive discrim­in­at­ory outcomes.

Worse, both the State Depart­ment policy and DHS proposal emerged from the Muslim ban and were preceded by Trump offi­cials’ state­ments that social media screen­ing was inten­ded to facil­it­ate ideo­lo­gical vetting aimed at Muslims. The genesis of these policies only under­scores how the drag­net collec­tion of social media iden­ti­fi­ers erects a digital screen­ing infra­struc­ture that is ripe for inten­tional, system­atic profil­ing.     

The Biden admin­is­tra­tion’s rejec­tion of the Trump-era DHS proposal to collect social media iden­ti­fi­ers on its travel and immig­ra­tion forms is a major victory for free expres­sion and privacy — both values enshrined in the U.S. Consti­tu­tion. The new admin­is­tra­tion should also roll back related social media surveil­lance programs, such as the State Depart­ment’s collec­tion program, that are groun­ded in the same flawed premises. Doing so would send an even stronger signal to the world that the United States remains commit­ted to the rights of people across the globe to speak and asso­ci­ate free of govern­ment scru­tiny.