For ICE/CBP the data regarding stops, searches, and arrests are thinner, more politically constrained, and more focused on patterns of Latino targeting, workplace/“street” arrests, and border stops rather than traffic stops generally.[1][2][3][4][5][6][7]

How racial profiling is measured for ICE/CBP

What the results show for ICE/CBP

Available analyses consistently point to heavy, often disproportionate targeting of Latinos and to profiling in street/workplace operations and jail‑based screening.[6][7][2][4][5][3]

• A study of ICE’s Criminal Alien Program in Irving, Texas found “strong evidence” that local police increased discretionary arrests of Hispanics for petty offenses (especially minor traffic violations) once they had around‑the‑clock access to ICE in the jail, with no similar increase for non‑Hispanics or in underlying crime.[3]
• Newly obtained 2025 ICE arrest data show that roughly one in five ICE arrests are of Latino people picked up on the street with no prior criminal record or removal order; the analyst concludes ICE is “deliberately targeting workers in heavily Latino jobs and neighborhoods,” often on thin or no individualized suspicion.[4]
• The same analysis notes that when ICE experimented with “consensual” work‑crew approaches, agents admitted that their profiling—based largely on appearance and job type—was accurate less than 10 percent of the time, meaning most people targeted turned out not to be the intended immigration violators.[4]
• Historical and contemporary reviews of Border Patrol practices document long‑standing reliance on racialized cues (skin color, language, proximity to the border) in roving and checkpoint stops; one American Immigration Council report describes Border Patrol as “steeped in institutional racism” and notes that racial profiling remains “a relevant factor” in official stop criteria.[2]
• Recent analyses of ICE workplace raids and large‑scale street operations show that many people arrested have no serious criminal record and that the vast majority are Latino, which, when compared to local demographics and crime patterns, is cited as evidence of racially targeted enforcement rather than neutral public‑safety priorities.[7][5][6][4]

Advocacy groups also emphasize that DOJ’s general ban on racial profiling includes explicit exemptions for CBP and some ICE border‑adjacent activities, giving these agencies wider legal latitude to consider race/ethnicity than local police, which shapes both practice and measurement.[9]

Where facial recognition fits in immigration enforcement

DHS (including ICE and CBP) uses facial recognition and other biometrics for identity checks at ports of entry, airport “biometric exit” systems, investigative work, and some mobile field operations.[10][11][12][13]

• Civil‑rights reports describe FRT as “racist, from how it was built to how it is used,” noting higher error rates for darker‑skinned people and its integration into ICE/CBP surveillance to “target and track immigrant families.”[11][10]
• The ACLU documents that ICE, CBP, and other DHS components used Palantir‑supported tech in 2017 to locate and arrest about 400 family members and caregivers of unaccompanied migrant children, and warns that combining similar analytics with facial recognition would allow mass identification and location of immigrants across the country.[10][11]
• Reporting and advocacy analyses argue that ICE and CBP are “aggressively adopting” face recognition as part of a “border is everywhere” approach, extending border‑style surveillance and racialized enforcement deep into the interior and onto city streets.[12][10]
• Congressional oversight letters in 2025 raised concerns specifically about ICE’s use of mobile facial‑recognition tools, highlighting the risk of misidentification, the lack of clear legal standards, and the disproportionate impact on communities of color and immigrants.[13]

Because immigration databases (e.g., visa, passport, and mugshot repositories) and ICE/CBP watchlists are already skewed toward non‑citizens and communities of color, plugging facial recognition into these systems tends to magnify existing biases—who is scanned, who is flagged, and who is apprehended.[11][12][13][10]

Impacts in the immigration context

The intersection of profiling and FRT in immigration enforcement has several concrete and systemic impacts.[6][7][13][2][9][10][4]

• Individual harms: Latinos and other people of color are more likely to be stopped, questioned about status, or swept into raids or street arrests on the basis of appearance, language, or workplace; adding facial recognition increases the risk of being misidentified as removable or as a target in an ICE operation.[12][13][2][9][10][4]