USCIS had no antifraud strategy. Parole sponsors used Elvis's SSN.
Summary
In September 2022, GAO found USCIS's Fraud Detection and National Security Directorate had no regular process for assessing fraud risk, no antifraud strategy, and no way to measure whether its antifraud work was even effective -- and made six recommendations. DHS agreed with all six. In December 2025, GAO reviewed one specific program built in the years since: humanitarian parole for 774,000 people from Cuba, Haiti, Nicaragua, Venezuela, and Ukraine. FDNS's own 2024 fraud analysis of 2.6 million sponsor applications found thousands using fake passports and the identities of dead Americans -- including Elvis Presley's Social Security number. This time, DHS refused two of GAO's three recommendations.
2022: the paperwork for catching fraud didn't exist yet
had run fraud risk assessments for three types of immigration benefits, but only because had told it to before -- there was no entity responsible for doing this on a regular schedule. FDNS had made changes to its antifraud activities over the years, but never built a strategy to decide where those changes should go or evaluated whether they worked. Even the basics were shaky: 's staffing model for FDNS investigators rested on assumptions that didn't match how the office actually operated, undermined further by inconsistent data entry that meant the model's own inputs weren't reliable.
's six recommendations covered the gaps directly: identify what actually drives FDNS's workload, fix the data entry, build a process for regular fraud risk assessments, build an antifraud strategy, evaluate whether antifraud activities work, and set performance targets to track it. concurred with every one.
View data as table
| 2022 report (6 of 6 concurred) | 100 |
|---|---|
| 2025 report (1 of 3 concurred) | 33.3 |
2025: the program that tested every one of those gaps
Between May 2022 and September 2024, granted supporter-based humanitarian parole to about 774,000 noncitizens -- 69% through the CHNV process for Cubans, Haitians, Nicaraguans, and Venezuelans; 30% through Uniting for Ukraine; and roughly 2% through family reunification parole. Each required a U.S.-based sponsor to file an application vouching for the beneficiary's financial support.
View data as table
| CHNV (Cuba, Haiti, Nicaragua, Venezuela) | 69 |
|---|---|
| Uniting for Ukraine | 30 |
| Family reunification parole | 2 |
By early 2024, FDNS officials had grown concerned enough about fraud in CHNV and Uniting for Ukraine that they analyzed all 2.6 million sponsor applications had received for the two programs -- not just a sample. They found more than 3,000 forms submitted with images of fraudulent U.S. passports, some altered to swap in real citizens' identities: one bore journalist Connie Chung's photo, another actor Cote de Pablo's. Other filers used Social Security numbers and other biographical details belonging to dead people -- including, in one case, Elvis Presley's. As of July 2024, more than 1,400 beneficiaries had already arrived in the U.S. on sponsor applications tied to a deceased identity.
Family reunification parole had its own version of the same problem. Some invitations to apply were mailed to petitioners who had since died; those invitations produced 728 sponsor applications filed under a deceased person's name. went on to confirm half of them anyway.
View data as table
| USCIS confirmed the application | 364 |
|---|---|
| Not confirmed | 364 |
also documented 'shell applications' -- sponsor filings submitted with fabricated or incomplete information for the sole purpose of collecting a fee from the prospective beneficiary, not to actually sponsor anyone. FDNS found the practice was widespread across both programs. In one case cited, a Ukrainian beneficiary paid about $5,000 to a filer for an application later confirmed. Because no statute bars a sponsor from charging a fee to file the paperwork, said it had no authority to act against sponsors who solicited or accepted payment for it.
and began taking some corrective steps once the fraud analysis came back -- automating address checks, blocking duplicate filings, referring suspected criminal activity to . suspended CHNV and Uniting for Ukraine in summer 2024, then suspended all three parole processes in January 2025. But found still had no internal control plan it could apply the next time it stands up a new benefit program -- the same kind of proactive, structured risk management its 2022 recommendations had asked to build for its existing programs.
made three recommendations this time: one that build that internal control plan, and two that formally assess and apply the lessons learned from the parole processes to its other operations. agreed only to the first. On the other two, said terminating the parole programs already solved the problem and declined to act further. disagreed, noting that the staffing and resource challenges the parole processes exposed weren't unique to parole and could recur anywhere stands up a program quickly.
The takeaway
- A framework gap became a real fraud problem. 's 2022 review found had no regular fraud-risk process. Its 2025 review found what that let through: 1,400+ arrivals tied to dead people's identities and thousands of forged-passport filings, discovered only after FDNS analyzed all 2.6 million applications itself.
- Some fraud went unpunished by design, not oversight. confirmed a Ukrainian applicant's sponsor after the beneficiary paid the sponsor $5,000 -- and said it lacked legal authority to act against fee-charging sponsors at all, since no statute prohibits it.
- accepted the diagnosis in 2022 and rejected most of the prescription in 2025. Every 2022 recommendation was concurred with; three years later, agreed only to build an internal control plan and refused two of 's three recommendations, arguing that ending the parole programs was fix enough. says the underlying gaps -- the ones that let 774,000 people through with inconsistent vetting -- could resurface in any future program builds quickly.
The 2022 findings are from -22-105328, 'U.S. Citizenship and Immigration Services: Additional Actions Needed to Manage Fraud Risks' (September 2022), read directly and in full via an archived copy. The 2025 findings are from -26-107433, 'Humanitarian Parole: Identified Fraud Risks in Parole Processes for Noncitizens and Should Assess Lessons Learned' (December 11, 2025), also read directly and in full. The two reports examine distinct facets of fraud-risk management -- the general framework in 2022, its real-world application (and failure) in a specific, large-scale program in 2025 -- about 3.2 years apart.
Sources(2) ▾
- U.S. Government Accountability Office, U.S. Citizenship and Immigration Services: Additional Actions Needed to Manage Fraud Risks (2022-09-19) — -22-105328, a report to congressional committees examining 's Fraud Detection and National Security Directorate's (FDNS) systemic processes -- the general fraud-risk-management-framework facet, before any specific program's fraud outcomes were assessed against it. Distinct from doc-gao26-107433 (a concrete case study of what happens when those systemic gaps play out in an actual program), issued about 3.2 years later. Direct gao.gov fetch returned HTTP 403; fetched via an existing Wayback capture. gao.gov · original document
- U.S. Government Accountability Office, Humanitarian Parole: DHS Identified Fraud Risks in Parole Processes for Noncitizens and Should Assess Lessons Learned (2025-12-11) — -26-107433, a report to congressional requesters examining 's supporter-based humanitarian parole processes -- the applied-consequence facet, showing how the framework gaps identified in doc-gao22-105328 manifested concretely in a specific, large-scale program. Fetched directly from files.gao.gov (HTTP 200); a Wayback capture was found to already exist at read time. gao.gov · original document
Comments
Always open. Logged-in readers can annotate paragraphs in place.
In September 2022⧉, found that 's Fraud Detection and National Security Directorate (FDNS) had no regular process for assessing fraud risk across the immigration benefits it adjudicates, no antifraud strategy to guide where it spent its resources, and no way to measure whether its antifraud work was effective. made six recommendations; agreed with all of them. In December 2025⧉, went looking at what that gap actually cost: a review of humanitarian parole programs that had, by then, let 774,000 people into the country found FDNS's own 2024 internal fraud analysis of 2.6 million sponsor applications -- more than 3,000 with fraudulent passport images, and identities lifted from dead Americans, including Elvis Presley's Social Security number.