The Perils of Facial Recognition in Law Enforcement

Flawed facial recognition technology used by law enforcement is causing wrongful arrests and threatening civil liberties, especially for people of color. Its potential for authoritarian abuse highlights the urgent need for widespread bans and regulations to protect communities.

The knock on the door, the sudden confrontation, the chilling assertion of authority – for Juan Carlos Lopez-Gomez, a U.S. citizen with a Social Security card in his pocket, it was a terrifying descent into a Kafkaesque nightmare.

On April 16, he found himself arrested on the unfounded suspicion of being an “unauthorized alien,” held for 30 harrowing hours by Immigration and Customs Enforcement (ICE).

His crime? Looking like someone else, thanks to a faulty biometric confirmation, a glaring error by facial recognition technology.

Lopez-Gomez’s ordeal is not an isolated incident.

Jensy Machado, another American citizen, was held at gunpoint and handcuffed by ICE agents, a victim of mistaken identity after someone else provided his home address on a deportation order.

These stories paint a stark picture of immigration policing in 2025: a world where suspicion trumps due process, where the arrest often precedes any real verification.

It’s a reality increasingly shaped by the embrace of error-prone technology, particularly facial recognition, by agencies like ICE.

The risks are escalating with alarming speed.

Due process safeguards, once sacrosanct, are being shredded, leaving citizens and noncitizens alike vulnerable to the arbitrary power of the state.

The very tools meant to enhance security are instead becoming instruments of uncertainty and potential tyranny.

Last month, it was revealed that Customs and Border Protection (CBP) sought proposals from tech firms to expand their use of an especially fallible facial recognition technology – precisely the kind that ensnared Lopez-Gomez.

Adding to the unease, ICE already holds nearly $9 million in contracts with Clearview AI, a facial recognition company that, despite its troubling ties to white nationalist groups, was once the most widely used private facial recognition system by federal agencies.

When reckless policing converges with powerful, inaccurate dragnet tools, the outcome is tragically predictable: more stories like those of Lopez-Gomez and Machado.

Studies have consistently demonstrated that facial recognition technology disproportionately misidentifies people of color, with Black women being particularly susceptible to errors.

In an era marked by a rapid surge in ICE activity, this technological flaw means more individuals are arbitrarily swept into the immigration dragnet, often without the fundamental right to due process to prove their legal standing.

The chilling truth is that even for American citizens who believe they have “nothing to hide,” merely resembling the wrong person can lead to jail time or even the terrifying prospect of deportation.

Yet, as dangerous as the technology’s mistakes are, its potential for abuse when it functions exactly as intended is even more chilling.

Imagine a presidential administration uncommonly willing to disregard legal procedures and judicial orders.

In such a climate, a perfectly accurate facial recognition system transforms into the most perilous possibility of all.

Federal agents could scour photos and footage of protests, using facial recognition to identify perceived political enemies.

These individuals could then be arrested, detained, or even deported, all without the basic protections of due process.

The very idea of the government wielding such a precise instrument for political retribution strikes at the heart of democratic principles.

The pervasiveness of facial recognition technology in our daily lives only amplifies this danger.

ICE, by collaborating with local law enforcement and private companies, and by sharing this technology, is expanding its capacity to round people up far beyond its traditional scope.

This deputization of surveillance infrastructure takes many forms: local police departments integrating facial recognition into their body cameras, landlords using it instead of keys to grant or deny tenants access, and stadiums employing it for security.

Until a recent moratorium, even New York public schools were using facial recognition on their security camera footage.

This creeping integration means that any facial recognition system, regardless of its initial benign intent, carries the inherent risk of being co-opted by ICE for indiscriminate or politically motivated deportations.

The good news is that resistance is taking root at state and local levels.

Cities like Boston, San Francisco, Portland, and the entire state of Vermont have already imposed regulations on facial recognition.

Florida has banned its use in schools, with Colorado poised to follow suit.

New York, ironically the most surveilled city in the country, has an opportunity to lead the charge with a package of “Ban the Scan” bills currently before its State Legislature and City Council.

These proposed laws aim to curb facial recognition use by police (S5609 / A1045), landlords (A6363 and Intro. 425), public accommodations (A6211 and Intro. 217), and to permanently ban its use in schools (S3827 / A6720).

While New York’s efforts are commendable, legislative protections against the profound risks of facial recognition must proliferate nationwide to truly safeguard our communities.

The ultimate defense against the specter of an authoritarian leader – be it the current one or any future administration – using this technology as a tool for political bidding lies in preventing its widespread use in the first place.

Given the immense power of facial recognition to target political adversaries and indiscriminately deport any citizen or noncitizen, and considering its racially biased inaccuracies are already causing wrongful arrests, a ban is not merely an option, but an essential step toward preventing authoritarian overreach.

The New York State Legislature has a clear path forward with the “Ban the Scan” package, and the rest of the country should heed its call, protecting communities from this force multiplier for tyranny.