The Trump administration’s sweeping overhaul of the Systematic Alien Verification for Entitlements program last year has sparked growing concerns about voter disenfranchisement, as evidence mounts that the expanded system is producing widespread errors in identifying eligible voters as potential noncitizens.

The changes, implemented in 2025 with the stated goal of eliminating what officials called the “taint” of illegal aliens defrauding Americans, fundamentally transformed how the SAVE program operates. Previously limited to individual searches for people with immigration records or formal Department of Homeland Security involvement, the program now allows bulk searches of anyone with a Social Security number. This enables election officials to query massive numbers of registered voters simultaneously.

The expansion also granted SAVE access to passport data from the Department of State, according to emails between U.S. Citizenship and Immigration Services and local election offices, as well as interviews with local officials. This represents a significant increase in the program’s scope and potential impact on voter registration maintenance.

While all states use data to maintain voter lists, large-scale data-matching programs inherently carry risks of false positives. Most election officials employ safeguards such as cross-referencing voter records against multiple state sources and providing voters opportunities to correct mistakes before removal from rolls. However, in previous years, some officials have attempted to remove voters without proper verification, jeopardizing eligible voters’ registration status and fueling unfounded doubts about election integrity.

That risk appears magnified in 2026 as the Trump administration uses the SAVE program to support claims that noncitizens are voting in large numbers. Three principal concerns have emerged regarding the reliability of voter fraud allegations based on SAVE results.

Large-scale data matching consistently produces errors across various contexts. Individuals often appear in multiple databases with inconsistencies in names, birthdates, or gender that prevent complete record queries. Data ecosystems also struggle to integrate new information, such as name changes after marriage. These gaps create problems for voter list maintenance, particularly when naturalized citizens have citizenship recorded in USCIS databases but not in Social Security Administration records.

USCIS maintains databases containing more than 100 million immigration records accessible through SAVE. Managing this volume is enormously complex, and even under optimal circumstances, mistakes occur when identifying citizenship or immigration status. A modest error rate at this scale could translate into tens of thousands of voters mistakenly identified as noncitizens.

Peer-reviewed social scientists consider false-positive error rates between 0.9 percent and 2.5 percent acceptable for large-scale data reconciliation. The New York Times reported earlier this year that DHS ran 49.5 million voter files through SAVE and identified around 10,000 registrants as potential noncitizens—just 0.02 percent of the total. This percentage falls far below typical error rates, suggesting the entire number could be explained by inherent matching errors rather than actual noncitizen registrations.

With approximately 174 million registered voters nationwide, even a small baseline error rate could create the false impression of massive numbers of ineligible voters on rolls. Americans falsely flagged could face unnecessary citizenship verification burdens or complete disenfranchisement if they miss notifications from election officials.

The program’s specific data sources compound reliability concerns. SAVE pulls from particularly error-prone datasets, including U.S. Customs and Border Protection’s Automated Targeting System and TECS, which incorporate information from the FBI’s terrorism watchlist. That watchlist contains unreliable information misidentifying people as terrorists and permits intentional biases against certain religious and racial minorities.

Newly accessible Social Security Administration data has significant shortcomings. The agency only began systematically collecting citizenship data in 1978, and citizenship notations for earlier Social Security number holders may have been inferred. The central database often lacks current citizenship information for naturalized citizens who failed to notify the agency of their naturalization. Additionally, roughly half of American citizens lack passports, making State Department passport data meaningless for tens of millions.

No dataset feeding SAVE contains up-to-date information on naturalized citizens. DHS publishes lists of hundreds of thousands of newly naturalized citizens quarterly, meaning substantial numbers of recently naturalized people may not have updated records when searches are conducted.

USCIS’s hasty expansion was finalized before citizenship data updates could be processed and merged, resulting in acknowledged blunders. The agency admitted providing incorrect information to at least five states. In Missouri, the program wrongly flagged hundreds of voters as noncitizens. One Boone County clerk reported that more than half of those flagged in November were actually U.S. citizens. An election official concluded the expanded SAVE program “is not ready for prime time.”

Real-world investigations have consistently disproven claims of substantial noncitizen voter registration. In St. Louis County, Missouri, approximately 35 percent of voters identified as noncitizens by SAVE were naturalized U.S. citizens who registered at their naturalization ceremonies, according to the county’s Republican election director. Texas officials failed to cross-check state records confirming citizenship after announcing hundreds of potential noncitizens identified by SAVE.

States taking more responsible approaches have confirmed the rarity of noncitizen voting. Utah’s complete voter roll review found zero instances of noncitizen voting. Louisiana’s review covering four decades identified only 79 potential noncitizens who voted out of an estimated 74 million ballots cast. Louisiana’s Republican secretary of state declared that noncitizens illegally registering or voting “is not a systemic problem in Louisiana.”

These outcomes, combined with inherent large-scale data matching pitfalls and flawed data sources, demand significant caution regarding future claims about noncitizens on voter rolls based on SAVE program results.