Since 2016, President Trump and his administrations have made non-citizen voting central to their misleading claims about election results. Even in elections the president won, he has claimed that non-citizen voters were the explanation for why his margin was not larger. 

Non-citizen voting is exceedingly rare, and it has never happened with enough frequency to swing the result of an election (this is especially true at the national level). In spite of that, DOJ and DHS have expanded an existing U.S. Citizenship and Immigration Services (USCIS) system, the Systematic Alien Verification for Entitlements (SAVE) system, so that it can be used to check the citizenship status of U.S. citizens without providing information on how this expansion was accomplished or the accuracy of the results. 

Some states have already started using SAVE to examine their own voter rolls for noncitizens. But the likeliest outcome of using this system for that purpose is not the removal of noncitizens from voter rolls. Instead, it is more likely to disqualify eligible voters, perpetuating a false narrative about the security of our elections. It’s worth noting, too, that in states where SAVE has been used, the system has affirmed once again that noncitizen voting is vanishingly rare. 

The changes this administration has made to the SAVE system are part of a larger effort to weaponize government data, sow disinformation and distrust in elections, suppress eligible voters, and tilt the 2026 elections before a single ballot is cast. 

What is the SAVE system?

USCIS maintains immigration status data through the SAVE system, which was originally designed to help state and local governments determine legal noncitizen residents’ eligibility for various government benefits (E.g., WIC, SNAP). Because it was built to bridge the gap between state benefit administrators and federal immigration data — not to identify citizens —  SAVE uses USCIS identifiers rather than Social Security numbers, and frequently cannot reach a definitive conclusion about an individual’s immigration status. In 2016, 18.7% of the queries made to SAVE (which is roughly 3.8 million) returned an inconclusive result as to an individual’s immigration status. 

In spite of these limitations, the federal government has allowed states to verify the citizenship of specific voters through the SAVE system since 2012. In 2025, though, DOGE and DHS shifted the tool’s purpose to allow it to be used not just to investigate individual voters, but to screen for noncitizens across voter rolls more generally by combining the information in the database with Social Security numbers. Using a system with such a high error rate could lead to voter rolls that are inaccurate and could disqualify eligible voters across the country. 

The real problems with SAVE 

In addition to its general unreliability, the likelihood that DHS would be able to match existing records perfectly with Social Security numbers is unlikely. Across 174 million registered voters, the likelihood for errors is quite high, and might mean that two people with the same name and date of birth could be confused for one another. What’s more, even if the information regarding a specific voter is accurate, it’s also possible that it could be out of date, and might falsely flag a voter as a noncitizen who has since become a naturalized citizen. 

The challenges with this system led multiple states to purge actual citizens from their voter rolls ahead of the 2024 election. Alabama, for example, inactivated 3,251 individuals, 30 who were identified “by cross-checking the voter rolls with a database of people who had been issued noncitizen identification numbers from the Department of Homeland Security.” Later, they acknowledged that at least 2,074 of those individuals were eligible voters. Texas cancelled the registrations of 6,500 voters and turned 2,000 over to the attorney general for investigation. However, reporting later revealed that the Secretary of State had only identified 581 of these voters over three years as noncitizens. 

SAVE could be used to interfere with the 2026 election 

Despite our elections being safe and accurate, reports indicate that the administration is planning to collect and review states’ voter files using the SAVE system to attempt to identify non-citizen voters. But this administration’s potential use of SAVE is not designed to solve a real problem. Non-citizen voting is exceedingly rare and has never come close to affecting an election result. States have run elections safely for decades and already have processes to keep voter rolls accurate. 

But an unreliable tool that removes thousands of voters is useful for something else: narrative. If the administration can claim (falsely) that SAVE identified thousands of noncitizens on voter rolls — without disclosing how many of those were eligible citizens wrongly flagged — it can use that claim to cast doubt on any election result it dislikes. The danger of the SAVE tool is not just that it is inaccurate, but that its inaccuracies could be used to perpetuate a broader narrative of election fraud, one that has no basis in actual fact. 

What We Can Do

The administration’s power here is more limited than it appears. States control their own elections — and without state voter files, SAVE cannot be weaponized against state voter rolls. That’s not a loophole. It’s a firewall, and states are already using it: filing suit to block SNAP data transfers, declining to share unredacted voter files, and restricting federal access to DMV databases. 

The administration is counting on the scale and complexity of this effort to feel overwhelming — but this isn’t inevitable. Free and fair elections in November 2026 won’t be won on Election Day only. They’ll be protected or attacked in the months before it — in the decisions states make now about what data they share and with whom. The work of protecting elections and voter data privacy is now, and some states are showing the way.