DHS 'Worst of the Worst' Website: Government Misinformation Exposed

CNN's February 19, 2026 investigation found hundreds of people listed as arrested for violent offenses who had only minor offenses on their records: single traffic violations, marijuana possession, or illegal reentry — a federal charge that involves crossing the border after a previous deportation, not a violent act. ^[1]

After receiving CNN's questions, DHS issued a spokesperson statement that same day. The statement admitted that charges against "hundreds" of individuals were "described incorrectly," attributed the cause to a "glitch," and claimed the affected entries represented "about 5%" of all entries — roughly 1,250 individuals out of 25,000 at the time. ^[3] Critically, the statement:

The complete absence of a public correction list — combined with immediate site expansion — means the public has no way to identify which of the 30,000+ current entries may contain original errors. Maine reporters, reviewing their state's own 13 listings, found at least one person whose crimes were committed decades earlier and out of state, one with no apparent felony conviction, and several who were transferred directly from local law enforcement into federal custody rather than arrested in the community. ^[14]

Representative Dan Goldman (D-NY) stated in a February 2026 interview: "The people they're going after are nowhere close to the worst of the worst. In fact, the vast majority of them have no criminal history whatsoever ... their statistics are completely inflated and made up." ^[12]

DHS's "5% glitch" admission — the charge mischaracterization CNN identified — is only one of three distinct error patterns visible across the independent reporting on wow.dhs.gov. A second pattern, documented by the Minnesota Department of Corrections, is custody-transfer mischaracterization: routine transfers from state prisons or county jails relabeled by DHS as new ICE "arrests." A third, visible in NPR's review of Minnesota-highlighted cases and in Maine reporters' review of their state's listings, is the use of decades-old, out-of-state, or unverifiable criminal history to make individuals appear more dangerous than the documented record supports. These three patterns are additive, not overlapping — and DHS has publicly addressed only the first. ^{[1] [5] [4] [14]}

The combined error rate across all three patterns is unknown. As of this report, no independent federal audit of the site's accuracy has been announced or conducted. ^{[1] [5]}

The case of Mauricio Morales-Morales is a clear documented instance of a DHS listing being factually wrong, not merely miscategorized. According to the Minnesota Department of Corrections, Morales-Morales was transferred out of state custody on June 2, 2025 for a court hearing on a judicial writ at federal authorities' request, then returned to DOC custody at the St. Cloud facility on September 24, 2025, where he has remained. DHS's site nonetheless describes him as arrested by federal agents and currently in ICE custody. ^[16]

This is not a "glitch" in how a charge was categorized — it is DHS publicly claiming custody of a man who, according to the state's own records, never left Minnesota state custody. The DOC's release documents 68 disputed cases of transfers mischaracterized as arrests, citing custody records and transfer logs, with video documentation included for several of the named examples. ^[16]

Other named cases in the DOC's January 26 release included Jose Eliborio Ocampo-Leon (Moose Lake facility, transferred to ICE January 22, 2026 — DHS claimed "arrest" the following day), Heydi Aguilera-Bustillo (Shakopee facility, transferred January 6, 2026, labeled as a federal enforcement operation), and Jaime Tirado-Hernandez (Lino Lakes facility, transferred January 8, 2026, same mischaracterization). In each case, the documentary record shows a routine administrative handoff — not an ICE arrest operation. ^[5]

DOC Commissioner Paul Schnell called DHS's public claims about the transfers "fundamentally false." ^[9]

Beyond the wow.dhs.gov site itself, DHS made a related statistical claim that Minnesota DOC was able to precisely refute with state records. DHS publicly claimed that 1,360 individuals with ICE detainers were being held in Minnesota custody. The Minnesota DOC conducted a systematic survey and found 301 total — 207 in state prisons and 94 in county jails. ^[9]

Commissioner Schnell said DHS provided no data source, methodology, or explanation for how the 1,360 figure was produced: "We cannot explain how those numbers square." ^[9]

DHS also suggested Minnesota was failing to cooperate with ICE detainers. Schnell disputed this, saying the DOC has cooperated with ICE detainers "as a matter of policy ... for a long, long time," and noting the state transferred 84 people directly from state prisons to ICE custody in 2025. ^{[9] [17]} This pattern — where DHS publishes a number roughly 4.5x the documented reality — mirrors the wow.dhs.gov pattern of inflating the scope of immigration enforcement for political messaging purposes.

Secretary Noem's exact claim: "70% of them have committed or have charges against them on violent crimes." PolitiFact's January 23, 2026 investigation found the accurate figure for violent crime convictions was approximately 5%, not 70%, citing a Cato Institute analysis; the New York Times separately found 7%. ^[19] The Cato Institute, using ICE booking data obtained via FOIA, found that 73% of individuals booked into ICE detention facilities since October 1, 2025 had no criminal convictions of any kind, and only 5% had violent criminal convictions. ^[20]

DHS publicly called the Cato analysis "made up" with "no legitimate data behind it." Newly released ICE booking data covering October 1–15, 2025 then confirmed Cato's findings: 71% of arrests in that window had no criminal convictions, and 45% had neither convictions nor pending charges — the same kind of data DHS had dismissed as fabricated. ^[21]

NPR published its analysis of 130 Minnesota arrest cases highlighted by DHS on social media on February 27, 2026. About one in seven involved convictions from at least 20 years earlier; six people had no criminal convictions at all and were highlighted based on pending charges; and for 37 people, NPR could not verify matching criminal histories in available records. The findings contradicted DHS's "worst of the worst" framing at scale. ^[4]

Juliet Stumpf, immigration law professor at Lewis & Clark Law School, told NPR: "I've never seen anything like this when it comes to immigration enforcement in the modern era." ^[4]

The site's geographic framing — searchable by state — was designed to enable localized messaging: "see who ICE removed from your community." Advocacy groups noted that public-facing criminal databases can damage reputations when corrections receive a fraction of the original story's reach. Immigration detainees facing removal proceedings have limited access to counsel, cannot easily contest public listings, and many lack the English-language capacity to alert media to errors. No reporting reviewed for this piece describes any DHS-established process by which a falsely listed individual can petition for removal or correction of their entry. ^[15]

Secretary Kristi Noem testified before the Senate Judiciary Committee on March 3, 2026 — her first appearance before the committee after declining prior requests. Senators from both parties pressed her on DHS's enforcement record; Sen. Thom Tillis (R-NC) told her "what we have seen is a disaster under your leadership," citing U.S. citizens who had been wrongly detained by immigration agents. Asked by Sen. Durbin whether she was wrong to call Alex Pretti — a VA nurse shot and killed by federal agents in Minneapolis — a "domestic terrorist," Noem declined to retract the characterization. ^[18]

The hearing followed the Minneapolis shootings of two U.S. citizens, Renee Good and Alex Pretti, by federal immigration agents in January 2026, and a House Oversight Committee Democratic staff report concluding that a "culture of impunity" at DHS "created heightened conditions for the use of deadly force against Americans." The report found that Trump administration officials lied to the public in the aftermath of the shootings. ^[22]

The wow.dhs.gov errors do not exist in isolation. They are part of a documented DHS pattern of public statements about immigration enforcement that other officials' own testimony has contradicted:

Testifying before the Senate Homeland Security Committee on February 12, 2026 — before the domestic-terrorist label was ever walked back by DHS leadership — ICE Acting Director Todd Lyons said, when asked whether he considered Pretti a domestic terrorist, "to my knowledge, no," and CBP leadership similarly declined to endorse the characterization used by Noem and White House aide Stephen Miller. ^[23] Senator Dick Durbin (Ranking Member), on Noem's delay in appearing: "Secretary Noem refused to appear before the Senate Judiciary Committee last year and now tells us that she will be available in five weeks — should she still be DHS Secretary at that time." ^[18]

The wow.dhs.gov website serves a specific political function that is structurally incompatible with accuracy. Understanding that function explains why errors are not random noise — they consistently make enforcement targets appear more dangerous than they are.

The 3,000-a-day arrest pressure: Trump administration officials, including Stephen Miller and Secretary Noem, publicly directed ICE to reach 3,000 arrests per day. ICE's historical capacity was roughly 300–310 arrests per day in the prior fiscal year. Hitting roughly 10x prior capacity in a compressed window required arresting people who were not violent criminals — but the site was designed to make all arrests appear to be violent criminal removals. CNN's analysis found that some of the cities with the highest apparent arrest concentrations on the site were small towns that coincidentally house large detention facilities (Conroe, Texas, home to the privately run Joe Corley Processing Center, topped the list) — suggesting many listed "arrests" were people already in federal custody or transferred from state sentences, not active community enforcement. ^[1]

The "criminal" definition inflation: By defining "criminal" to include pending charges, decades-old minor convictions, and immigration civil violations themselves, DHS built a self-confirming narrative: everyone on the site is a criminal because DHS says they are. The distinction matters legally and constitutionally — a pending charge means a person has been accused, not convicted. The presumption of innocence is a constitutional guarantee. ^[19]

No disclosed correction mechanism: No reporting reviewed for this piece describes any DHS-established process by which an individual falsely listed on the site can petition for removal or correction. Courts have found that government-operated registries without correction mechanisms can raise due-process concerns when they carry reputational and legal consequences.

The ICE arrest data trend: ICE booking data show the share of detainees with criminal convictions fell from 44.7% in the first three months of Trump's second term to 31.8% in the three months ending mid-October 2025. Only about 7% had violent convictions, per a New York Times analysis. Nearly 43% had no U.S. criminal record at all as of January 2026 — up from 21.9% in the administration's first three months. That rise directly contradicts the "worst of the worst" branding. ^[7]

The wow.dhs.gov website was described by DHS as a transparency and accountability tool — a way for citizens to see who their government was removing from their communities. The documented record shows something different: a government-operated database that published false criminal profiles of identifiable individuals, was used to politically justify mass enforcement operations conducted amid fatal shootings of U.S. citizens, and was expanded after its errors were proven rather than corrected.

The pattern documented across DHS's communications follows a consistent structure: DHS makes a public claim with the authority of a government seal; independent investigators or state officials produce documentary evidence disproving it; DHS does not retract; DHS escalates its activities in the same area. This happened with the glitch claim, the Minnesota arrest numbers, the ICE detainer count, the Cato Institute data, the "domestic terrorist" label for Alex Pretti, and a White House graphic that falsely accused Victor Manuel Carranza — whose actual convictions were for larceny and identity theft — of child sex crimes. ^{[6] [21] [22]}

Who is harmed: Individuals falsely labeled — their names and mugshots appear on a searchable federal government website linked to crimes they did not commit. Corrections receive a fraction of the original story's reach. Some individuals are detained during the legal process and cannot dispute the listing. Two U.S. citizens are dead in Minneapolis. State officials who contradict DHS publicly with documentary evidence receive far less social media amplification than DHS's original false claims. ^{[4] [15]}

The single most important fact about wow.dhs.gov: DHS has never published a list of which entries contain errors. After admitting that "hundreds" were described incorrectly, the agency added 5,000 more entries and issued a press release expressing gratitude to CNN for publicizing the site. A person searching the database for an individual they know has no way to determine whether that listing is accurate, partially accurate, or one of the entries DHS privately knows is wrong.