The Trump administration has launched an AI-driven review of more than 55 million people holding U.S. visas as part of a program the State Department described as "continuous vetting," officials and former employees told the Daily Mail. The initiative, officials say, is intended to identify visa overstays and other potential bases for revocation, including criminal activity and terrorism-related concerns.
Sources familiar with the review told the Daily Mail that the State Department is turning to automated, AI-powered tools to cross-check records, social media accounts and partner-country immigration databases. A former State Department employee told the newspaper that the broad framing of the sweep could also operate as a form of deterrence, encouraging people who believe they are ineligible to leave the country voluntarily rather than face federal enforcement.
The State Department confirmed to the Daily Mail that all U.S. visa holders will face "continuous vetting" and said available information would be used to identify potential reasons to bar admission or continue living in the United States. The review comes amid recent policy moves including tighter student visa restrictions and a roughly 20% staff reduction at the department, a combination officials and former employees say has increased reliance on technology.
Immigration experts and technologists expressed concerns about accuracy, transparency and the potential for wrongful revocations. Julia Gelatt, associate director of the U.S. Immigration Policy Program at the Migration Policy Institute, told the Daily Mail she wants more public information about how the system will prioritize cases and what databases will be used. Gelatt said some government data sources are incomplete and that reliance on automated matches could lead to incorrect outcomes, for example if an arrest that did not result in conviction is recorded in a shared database.
Technology analyst Rob Enderle, president and principal analyst at the Enderle Group, warned that prioritizing speed and productivity in automated systems can reduce accuracy. "There is a far greater focus on productivity than quality," he said, adding that the likelihood of errors or misclassifications is "exceptionally high" unless automated outputs are paired with extensive human review and testing.
Advocates and lawyers pointed to recent, high-profile cases that critics say illustrate the risks of automated flagging. In April, Japanese student Suguru Onda, a student at Brigham Young University, had his visa terminated over a fishing citation and traffic tickets, a situation his attorney called likely the result of an AI error. In March, Turkish student Rümeysa Öztürk was arrested after her F-1 visa was revoked and transferred to ICE custody, a move that drew criticism from lawmakers and civil-rights groups.
A State Department official quoted by Fox News defended revocations in student cases, saying they were tied to lawbreaking or support for terrorism. The Daily Mail account and current and former employees, however, said officials are likely to prioritize certain countries and populations even if the department publicly frames the effort as a broad sweep.
Officials familiar with the program told the Daily Mail the department has already used AI-supported systems in recent student visa actions and expects to draw selectively from the larger pool of visa holders. Gelatt said the 55 million figure is misleading because many people counted do not live in the United States and because the practical focus will be narrower, resembling an ongoing database check such as ICE's continuously monitored data center.
The Department of Homeland Security reported nearly 13 million lawful permanent residents and roughly 4 million people on temporary visas in the United States last year, figures experts say underscore the scope and complexity of any automated review.
Civil-rights groups and immigration advocates have called for transparency, independent testing and safeguards to ensure that automated tools do not produce systemic errors or politically motivated targeting. Enderle recommended extensive validation and continued human oversight until error rates reach an acceptable level; Gelatt urged clearer explanations about data sources, prioritization criteria and redress mechanisms for those flagged by automated systems.
The administration has not publicly released a detailed plan for how AI will be used across the visa population or what safeguards will exist for contested cases. The State Department and Homeland Security officials who discussed the program with the Daily Mail characterized the effort as part of routine national-security and immigration enforcement work.
As the review proceeds, legal advocates, technologists and policy experts said they will be watching for evidence of testing, human review, and avenues for individuals to contest automated findings, noting that deployment at scale without those safeguards could produce wrongful revocations and raise novel enforcement and civil-rights questions.