Government Surveillance, Privacy & Data
State Department Should Abandon Its Plan to Collect Social Media Information From 14.7 Million Visa Applicants
In March the State Department (DOS) issued a notice proposing that all immigrant and nonimmigrant visa applicants be required to provide social media identifiers, used in the past five years and from a list of 20 platforms, “for identity resolution and vetting purposes based on statutory visa eligibility standards.” This collection request would have an immediate impact on 14.7 million visa applicants, and thousands, if not millions, more third parties whose data could be collaterally reviewed. DOS has asserted—without support—that evaluating an applicant’s eligibility for a visa “would not be possible without collecting this information.” DOS issued a version of this request in May 2017, targeted to a subset of visa applicants, which was opposed by CDT and many other human rights and civil liberties groups.
On Tuesday, CDT filed comments opposing this latest social media information request and joined a coalition comment with 56 other civil liberties and human rights organizations. In these comments, we highlighted the following problems with DOS’s proposal:
The notice is vague and overbroad and will lead to unintentionally incomplete applications and adverse determinations.
The requirement to report social media identifiers for the listed social media platforms used in the last five years scopes too broadly: It is not reasonable to expect a visa applicant to remember all of the social media platforms they have used in the last five years. The average person has 7.6 active social media accounts and may have temporarily created many more to test out new platforms. Visa applicants can also easily overlook or forget that they own certain accounts, particularly when they use their email or Facebook and Twitter account to log into a new platform, or when a profile is automatically created for them. It is also unclear how much activity on a platform would trigger mandated disclosure, and what applicants would have to disclose if they use platforms without a profile.
This proposal lacks guidance about how social media content will be collected, retained or shared. Moreover, it fails to address the consequences applicants would suffer from failing to disclose every social media platform they use and their various identifiers. Would such a mistake be grounds for visa denial, or future revocation of an immigration benefit? Given the realities of how people adopt and discard social media accounts, and the high stakes for visa applicants, immigration officials should be presumptively lenient about such mistakes, particularly as visa applicants are not afforded any path to appeal an adverse decision.
The consequences of the vagueness and overbreadth of this proposal are stark: the U.S. will lose visa applicants from the very people the U.S. seeks to attract. A prohibitively difficult or unattractive barrier to entry causes the U.S. to lose its competitive edge in education, business, and tourism.
Social media content is not easily interpreted and will likely lead to the use of problematic algorithmic screening.
Social media communication, like most human interactions, is idiosyncratic. Deciphering the meaning of statements is difficult without an intimate understanding of the context in which they are made. Parsing meaning from text is particularly difficult when communications employ slang, sarcasm, or non-textual information including emojis, GIFs, and “likes.” Visa applicants’ social media content will also often contain foreign languages, further increasing the complexity of analyzing this information.
Moreover, DOS’s intent to access vast amounts of social media information, combined with its lack of a stated plan for using or parsing that information, raises the concern that DOS could attempt to implement an automated vetting system. Such a system would be inherently technologically deficient and prone to discrimination. As CDT has explained, automated tools for analyzing the text of social media posts cannot reliably interpret the meaning of a post or the speaker’s intent. Because automated tools cannot conduct a nuanced evaluation of an applicant for entry into the United States, DOS would likely resort to proxies that are easier for computers to discern, such as the existence of religious imagery in photos. Such proxies would likely facilitate discriminatory targeting.
Social media screening is unlikely to yield relevant security information.
DOS appears to believe that it can identify potential security threats by scrutinizing online speech, but research shows that such expressive conduct is not a valid predictor of one’s propensity to commit an act of violence. Furthermore, DOS has not demonstrated that social media data will provide it with the ability to better enforce the Immigration and Nationality Act. Rather, DOS already has access to ample resources to assess an individual’s admissibility for a visa. Additionally, there is no evidence to suggest that DOS has developed a strategy for evaluating social media data. DOS has solicited social media data for a year, and no audit has been conducted to demonstrate efficacy in that smaller collection. Finally, social media screening is comically easy for bad actors to circumvent. Knowing that DOS will be combing through social media data, would-be criminals and terrorists can simply delete or manipulate their online social media behaviors, and disclose only newly-created, sanitized social media accounts to DOS during the visa application process in order to evade detection.
DOS’s proposal will chill free speech, association, and inhibit the right to anonymity.
Knowing that their social media content will or could be monitored, immigrants and the individuals that interact with them will feel pressure to self-censor, delete their social media accounts, and disengage from online spaces. This policy will also chill the right to anonymous speech for applicants and U.S. persons, leading to negative repercussions for members of vulnerable communities such as domestic violence victims and religious minorities. Due to legitimate privacy and security concerns, individuals often use anonymous social media profiles, alter their online profiles from their real-world identities, or list conflicting information across their social media handles.
DOS’s collection proposal comes on the heels of multiple government efforts to monitor immigrants’ digital lives, including: Immigration and Customs Enforcement’s Extreme Vetting Initiative, DOS’s aforementioned existing social media collection, Customs and Border Protection’s searches of digital devices at the border, and the Department of Homeland Security’s recent decision to retain social media data in Alien Files. CDT has consistently worked with coalition partners to challenge these programs, and will continue to oppose the government’s efforts to collect, retain, share and use social media data.