Skip to content 
U.S. Department of Homeland Security says it has corrected a public-facing database that the agency and the White House have repeatedly promoted as evidence of a crackdown on what officials call “the worst of the worst” immigrants, after questions about why the site appeared to misstate or minimize alleged offenses for hundreds of people.
The database, branded “Arrested: Worst of the Worst,” lists roughly 25,000 individuals along with the crimes the agency says they are tied to. In statements reported by local outlets citing the CNN review, DHS acknowledged that entries for hundreds of people were described incorrectly and blamed the problem on a “glitch” it said was fixed by February 18, 2026.
A Public List Marketed as Transparency, Undermined by Basic Data Problems
View this post on Instagram
DHS has framed the website as a transparency tool showing the public who immigration authorities are arresting.
But the reporting around the site has centered on a straightforward problem: the information presented about individual cases did not reliably match the “worst of the worst” framing.
The CNN review, echoed in subsequent coverage, found the site had numerous inaccuracies in how it displayed criminal allegations, at times appearing to link people to minor offenses such as traffic-related violations or marijuana possession while the government’s public messaging implied a roster dominated by violent criminals.
DHS responded that the mislabeling stemmed from a technical issue and said many listed under minor offenses had additional crimes not correctly shown.
DHS did not publicly detail what kind of technical failure could misdescribe charges across hundreds of entries, according to the same reporting.
Why the Location Data Raised Separate Questions
Beyond offense categories, the site’s “city of arrest” field has drawn scrutiny for what it may imply about where arrests are happening.
Reporting tied to the CNN review highlighted that some of the most common arrest locations on the list are smaller jurisdictions known for major detention facilities, creating the possibility that the public-facing database reflects transfers or custody changes rather than arrests originating in neighborhoods.
State officials in Minnesota have also accused federal authorities of inflating or muddying arrest narratives by taking credit for people initially arrested by local law enforcement and later transferred through routine processes.
A Minnesota corrections official, quoted in reporting on the database, said inconsistencies in DHS data raised questions about how federal resources were being applied to the “worst of the worst” targets in the state.
The Site Lands in the Middle of a Bigger Debate Over Who Is Being Arrested
The controversy also intersects with a broader body of reporting and data about the composition of immigration arrests and detention.
An internal DHS document obtained by CBS News found that less than 14% of nearly 400,000 people arrested by U.S. Immigration and Customs Enforcement during President Donald Trump’s first year back in office had charges or convictions for violent criminal offenses.
The same reporting said ICE made roughly 393,000 arrests between January 21, 2025 and January 31, 2026, and that nearly 40% of arrestees had no criminal record and were accused only of civil immigration violations.
Independent analysis has pointed to similar patterns. Migration Policy Institute reported that as of January 7, 2026, 26% of people in ICE detention had a criminal conviction, 26% had a pending criminal charge, and 48% had an immigration-related charge.
A separate Reuters investigation has documented the legal pressure building around detention practices, reporting that federal courts have ruled thousands of times since October that ICE was unlawfully detaining immigrants.
Reuters also reported that the number of people in ICE detention reached about 68,000 in February 2026, roughly 75% higher than when Trump took office the prior year.
Communications Risk for DHS: One Label, Many Different Cases
DHS has used the “worst of the worst” label as a catch-all political argument for aggressive enforcement. A public database with faulty or unclear offense descriptions creates a predictable vulnerability: it invites case-by-case scrutiny that can undercut a message built for mass effect.
That dynamic is part of what other coverage has emphasized. A report distributed by WABE and credited to NPR described a broader pattern of high-profile DHS claims that were later challenged by evidence, dropped charges, or judicial findings, with critics arguing the rhetoric often gets ahead of verified facts.
For DHS, the immediate question is narrow: what exactly went wrong with the “Worst of the Worst” site and how the agency validated the correction.
For the public, the larger question remains structural: whether a database marketed as proof of precision targeting can earn credibility when the underlying data presentation proves fragile.
