DOGE Is Building a Master Database to Surveil and Track Immigrants

Operatives from Elon Musk’s so-called Department of Government Efficiency (DOGE) are building a master database at the Department of Homeland Security (DHS) that could track and surveil undocumented immigrants, two sources with direct knowledge tell WIRED.

DOGE is knitting together immigration databases from across DHS and uploading data from outside agencies including the Social Security Administration, as well as voting records, sources say. This, experts tell WIRED, could create a system that could later be searched to identify and surveil immigrants.

The scale at which DOGE is seeking to interconnect data, including sensitive biometric data, has never been done before, raising alarms with experts who fear it may lead to disastrous privacy violations for citizens, certified foreign workers, and undocumented immigrants.

A United States Customs and Immigration Services (USCIS) data lake, or centralized repository, existed at DHS prior to DOGE that included data related to immigration cases, like requests for benefits, supporting evidence in immigration cases, and whether an application has been received and is pending, approved, or denied. Since at least mid-March, however, DOGE has been uploading mass amounts of data to this preexisting USCIS data lake, including data from the Internal Revenue Service (IRS), SSA, and voting data from Pennsylvania and Florida, two DHS sources with direct knowledge tell WIRED.

“They are trying to amass a huge amount of data,” a senior DHS official tells WIRED. “It has nothing to do with finding fraud or wasteful spending … They are already cross-referencing immigration with SSA and IRS as well as voter data.”

Since president Donald Trump’s return to the White House earlier this year, WIRED and other outlets have reported extensively on DOGE’s attempts to gain unprecedented access to government data, but until recently, little has been publicly known about the purpose of such requests or how they would be processed. Reporting from The New York Times and The Washington Post has made clear that one aim is to cross-reference datasets and leverage access to sensitive SSA systems to effectively cut immigrants off from participating in the economy, which the administration hopes would force them to leave the county. The scope of DOGE’s efforts to support the Trump administration’s immigration crackdown appear to be far broader than this, though. Among other things, it seems to involve centralizing immigrant-related data from across the government to surveil, geolocate, and track targeted immigrants in near real time.

DHS and the White House did not immediately respond to requests for comment.

DOGE’s collection of personal data on immigrants around the US has dovetailed with the Trump administration’s continued immigration crackdown. “Our administration will not rest until every single violent illegal alien is removed from our country,” Karoline Leavitt, White House press secretary, said in a press conference on Tuesday.

On Thursday, Gerald Connolly, a Democrat from Virginia and ranking member on the House Oversight Committee, sent a letter to the SSA office of the inspector general stating that representatives have spoken with an agency whistleblower who has warned them that DOGE was building a “master database” containing SSA, IRS, and HHS data.

“The committee is in possession of multiple verifiable reports showing that DOGE has exfiltrated sensitive government data across agencies for unknown purposes,” a senior oversight committee aide claims to WIRED. “Also concerning, a pattern of technical malfeasance has emerged, showing these DOGE staffers are not abiding by our nation’s privacy and cybersecurity laws and their actions are more in line with tactics used by adversaries waging an attack on US government systems. They are using excessive and unprecedented system access to intentionally cover their tracks and avoid oversight so they can creep on Americans' data from the shadows.”

“There's a reason these systems are siloed,” says Victoria Noble, a staff attorney at the Electronic Frontier Foundation. “When you put all of an agency's data into a central repository that everyone within an agency or even other agencies can access, you end up dramatically increasing the risk that this information will be accessed by people who don't need it and are using it for improper reasons or repressive goals, to weaponize the information, use it against people they dislike, dissidents, surveil immigrants or other groups.”

One of DOGE’s primary hurdles to creating a searchable data lake has been obtaining access to agency data. Even within an agency like DHS, there are several disparate pools of data across ICE, USCIS, Customs and Border Protection (CBP), and Homeland Security Investigations (HSI). Though some access is shared, particularly for law enforcement purposes, these pools have not historically been commingled by default because the data is only meant to be used for specific purposes, experts tell WIRED. ICE and HSI, for instance, are law enforcement bodies, and sometimes need court orders to access an individual's information for criminal investigations, whereas USCIS collects sensitive information as part of the regular course of issuing visas and green cards.

DOGE operatives Edward Coristine, Kyle Schutt, Aram Moghaddassi, and Payton Rehling have already been granted access to systems at USCIS, FedScoop reported earlier this month. The USCIS databases contain information on refugees and asylum seekers and possibly data on green card holders, naturalized US citizens, and Deferred Action for Childhood Arrivals (DACA) recipients, a DHS source familiar tells WIRED.

DOGE wants to upload information to the data lake from myUSCIS, the online portal where immigrants can file petitions, communicate with USCIS, view their application history, and respond to requests for evidence supporting their case, two DHS sources with direct knowledge tell WIRED. In combination with IP address information from immigrants that sources tell WIRED that DOGE also wants, this data could be used to aid in geolocating undocumented immigrants, experts say.

Voting data, at least from Pennsylvania and Florida, appears to also have also been uploaded to the USCIS data lake. In the case of Pennsylvania, two DHS sources tell WIRED that it is being joined with biometric data from USCIS’s Customer Profile Management System, identified on the DHS’s website as a “person-centric repository of biometric and associated biographic information provided by applicants, petitioners, requestors, and beneficiaries” who have been “issued a secure card or travel document identifying the receipt of an immigration benefit.”

“DHS, for good reason, has always been very careful about sharing data,” says a former DHS staff member who spoke to WIRED on the condition of anonymity because they were not authorized to speak to the press. “Seeing this change is very jarring. The systemization of it all is what gets scary, in my opinion, because it could allow the government to go after real or perceived enemies or ‘aliens; ‘enemy aliens.’”

While government agencies frequently share data, this process is documented and limited to specific purposes, according to experts. Still, the consolidation appears to have administration buy-in: On March 20, President Trump signed an executive order requiring all federal agencies to facilitate “both the intra- and inter-agency sharing and consolidation of unclassified agency records.” DOGE officials and Trump administration agency leaders have also suggested centralizing all government data into one single repository. “As you think about the future of AI, in order to think about using any of these tools at scale, we gotta get our data in one place," General Services Administration acting administrator Stephen Ehikian said in a town hall meeting on March 20. In an interview with Fox News in March, AirBnb cofounder and DOGE member Joe Gebbia asserted that this kind of data sharing would create an “Apple-like store experience” of government services.

According to the former staffer, it was historically “extremely hard” to get access to data that DHS already owned across its different departments. A combined data lake would “represent significant departure in data norms and policies.” But, they say, “it’s easier to do this with data that DHS controls” than to try to combine it with sensitive data from other agencies, because accessing data from other agencies can have even more barriers.

That hasn’t stopped DOGE operatives from spending the last few months requesting access to immigration information that was, until recently, siloed across different government agencies. According to documents filed in the American Federation of State, County and Municipal Employees, AFL-CIO v. Social Security Administration lawsuit on March 15, members of DOGE who were stationed at SSA requested access to the USCIS database, SAVE, a system for local and state governments, as well as the federal government, to verify a person’s immigration status.

According to two DHS sources with direct knowledge, the SSA data was uploaded to the USCIS system on March 24, only nine days after DOGE received access to SSA’s sensitive government data systems. An SSA source familiar tells WIRED that the types of information are consistent with the agency's Numident database, which is the file of information contained in a social security number application. The Numident record would include a person’s social security number, full names, birthdates, citizenship, race, ethnicity, sex, mother’s maiden name, an alien number, and more.

Oversight for the protection of this data also appears to now be more limited. In March, DHS announced cuts to the Office for Civil Rights and Civil Liberties (CRCL), the Office of the Immigration Detention Ombudsman, and the Office of the Citizenship and Immigration Services Ombudsman, all key offices that were significant guards against misuse of data. “We didn't make a move in the data world without talking to the CRCL,” says the former DHS employee.

CRCL, which investigates possible rights abuses by DHS and whose creation was mandated by Congress, had been a particular target of DOGE. According to ProPublica, in a February meeting with the CRCL team, Schutt said, “This whole program sounds like money laundering.”

Schutt did not immediately respond to a request for comment.

Musk loyalists and DOGE operatives have spoken at length about parsing government data to find instances of supposed illegal immigration. Antonio Gracias, who according to Politico is leading DOGE’s “immigration task force,” told Fox and Friends that DOGE was looking at voter data as it relates to undocumented immigrants. “Just because we were curious, we then looked to see if they were on the voter rolls,” he said. “And we found in a handful of cooperative states that there were thousands of them on the voter rolls and that many of them had voted.” (Very few noncitizens voted in the 2024 election and naturalized immigrants were more likely to vote Republican.) Gracias is also part of the DOGE team at SSA and founded the investment firm Valor Equity Partners. He also worked with Musk for many years at Tesla, and helped the centibillionaire take the company public.

“As part of their fixation on this conspiracy theory that undocumented people are voting, they're also pulling in tens of thousands, millions of US citizens who did nothing more than vote or file for Social Security benefits,” Cody Venzke, a senior policy counsel at the American Civil Liberties Union focused on privacy and surveillance, tells WIRED. “It's a massive dragnet that's going to have all sorts of downstream consequences for not just undocumented people, but US citizens and people who are entitled to be here as well.”

Over the last few weeks, DOGE leadership within the IRS have orchestrated a “hackathon” aimed at plotting out a “mega API” allowing privileged users to view all agency data from a central access point. Sources tell WIRED the project will likely be hosted on Foundry, software developed by Palantir, a company cofounded by Musk ally and billionaire tech investor Peter Thiel. An API is an application programming interface that allows different software systems to exchange data. While the Treasury Department has denied the existence of a contract for this work, IRS engineers were invited to another three-day “training and building session” on the project located at Palantir’s Georgetown offices in Washington, DC, this week, according to a document viewed by WIRED.

“Building it out as a series of APIs they can connect to is more feasible and quicker than putting all the data in a single place, which is probably what they really want,” one SSA source tells WIRED.

On April 5, DHS struck an agreement with the IRS to use tax data to search for more than seven million migrants working and living in the US. ICE has also recently paid Palantir millions of dollars to update and modify an ICE database focused on tracking down immigrants, 404 Media reported.

Multiple current and former government IT sources tell WIRED that it would be easy to connect the IRS's Palantir system with the ICE system at DHS, allowing users to query data from both systems simultaneously. A system like the one being created at the IRS with Palantir could enable near-instantaneous access to tax information for use by DHS and immigration enforcement. It could also be leveraged to share and query data from different agencies as well, including immigration data from DHS. Other DHS sub-agencies, like USCIS, use Databricks software to organize and search its data, but these could be connected to outside Foundry instances simply as well, experts say. Last month, Palantir and Databricks struck a deal making the two software platforms more interoperable.

“I think it's hard to overstate what a significant departure this is and the reshaping of longstanding norms and expectations that people have about what the government does with their data,” says Elizabeth Laird, director of equity in civic technology at the Center for Democracy and Technology, who noted that agencies trying to match different datasets can also lead to errors. “You have false positives and you have false negatives. But in this case, you know, a false positive where you're saying someone should be targeted for deportation.”

Mistakes in the context of immigration can have devastating consequences: In March, authorities arrested and deported Kilmar Abrego Garcia, a Salvadoran national, due to, the Trump administration says, “an administrative error.” Still, the administration has refused to bring Abrego Garcia back, defying a Supreme Court ruling.

“The ultimate concern is a panopticon of a single federal database with everything that the government knows about every single person in this country,” Venzke says. “What we are seeing is likely the first step in creating that centralized dossier on everyone in this country.”