The Trump administration is deploying artificial intelligence to carry out a review of more than 55 million people who hold U.S. visas, launching what officials characterize as a program of “continuous vetting” that could identify overstays, criminal activity or terrorist-related risks.
The State Department said the new process will examine information across multiple government databases, social media accounts and records from visa holders’ countries of origin to flag potential violations that could lead to revocation or removal. Administration officials have framed the effort as an upgrade to existing monitoring systems; critics and former employees say it could amount to the largest immigration sweep in U.S. history and raise serious accuracy and civil liberties concerns.
A former State Department official who spoke to the Daily Mail described the initiative as a psychological tactic aimed at encouraging people who have overstayed or otherwise lost eligibility to self-deport, saying the administration need not review all 55 million records in detail to achieve that effect. The official said the department will likely prioritize certain nationalities but cannot publicly acknowledge that targeting.
Migration Policy Institute Associate Director Julia Gelatt urged greater transparency about how the process will operate and how erroneous data will be addressed. “Different government databases are speaking to each other looking for matches, but there are concerns some have incomplete information — like FBI data — so if somebody has an arrest but is ultimately found innocent, that might not be recorded,” Gelatt said, according to reporting on the plan. She warned that visas could be wrongly revoked based on faulty records or misinterpreted online content.
Technology analysts and civil liberties groups have repeatedly cautioned about automated systems that examine large populations. Rob Enderle, president and principal analyst at the Enderle Group, said the odds of AI-driven reviews producing harmful errors are high if the tools are prioritized for speed over accuracy. “There is a far greater focus on productivity than quality,” he said, adding that automated flags should be subject to extensive testing and human review before leading to enforcement actions.
The plan arrives amid other enforcement moves by the administration, including newly tightened student visa policies that officials say have resulted in the revocation of roughly 6,000 student visas since January. Critics point to individual cases as evidence of the risks posed by automated or cursory reviews. In April, a Japanese student at Brigham Young University, Suguru Onda, had a visa terminated after what his attorneys say was an AI-related error tied to a fishing citation and traffic tickets. Separately, the arrest of Rümeysa Öztürk, a Turkish graduate student whose F-1 visa was revoked and whose detention drew criticism from lawmakers and civil rights advocates, has been cited as an example of aggressive enforcement.
Officials have defended the broader approach as a necessary modernization of visa oversight following staff reductions at the State Department. The department has cut consular staffing by about 20 percent, according to former officials, a change that proponents say makes automated tools more necessary to process and monitor large volumes of records. The same officials say the operation will likely triage leads to focus finite resources on higher-risk cases.
Data from the Department of Homeland Security shows there were nearly 13 million green-card holders and about 4 million people in the United States on temporary visas in the previous year, underscoring the scope of records that could be drawn into cross-checks even if only subsets are prioritized for enforcement. Experts say that without adequate human oversight, automated checks across interagency databases and social media could generate false positives that trigger visa revocations or deportation proceedings.
Advocates and immigration scholars called for clearer rules, independent oversight and mechanisms to challenge findings produced by automated systems. Gelatt said an ongoing database check resembling current ICE monitoring could be the realistic outcome rather than a literal manual review of every visa file, but she added that policymakers must specify standards for accuracy, appeals and privacy protections.
The administration has not publicly detailed the algorithms, vendor arrangements, or safeguards that will govern the program. As the government moves forward with the technical rollout, lawmakers, civil rights organizations and former officials say they will watch for how flagged cases are handled and whether automated findings are verified before enforcement actions are taken.