A Dutch activist claims he has developed AI software that can reveal hidden ICE agents’ identities.
According to Politico, Dominick Skinner says he has publicly identified at least 20 ICE agents during arrests. His algorithm can rebuild their faces from just 35% of visible facial features.
Since Donald Trump’s return to the White House, ICE agents have started wearing masks during migrant arrests. First, ICE points to safety concerns for this move. Then, officials say this step helps protect officers from both physical and digital retaliation.
Yet, critics see these masks as a sign of unchecked government power. Moreover, Skinner’s project, which uses AI facial recognition to share suspected identities online, adds to a bigger debate. This debate focuses on openness, accountability, and privacy rights.
“ICE agents shouldn’t be hunted online by activists using artificial intelligence,” says Republican Senator James Lankford from Oklahoma. As chair of the Homeland Security subcommittee, he shared his concerns with Politico.
Furthermore, Skinner’s work is part of a larger campaign called ICE List. This list shares names of hundreds of agents and office workers from the federal agency. As a result, both media outlets and the Department of Homeland Security (DHS) are paying attention.
While ICE hasn’t verified these identifications, their spokesperson Tanya Roman strongly responds: “These masks are for safety, not hiding. Such posts put agents’ lives at risk.”
Additionally, DHS views this project as doxxing – the harmful sharing of personal details. In fact, they believe this practice puts federal agents in serious danger.
Moving Toward Legislative Response
The rewritten version maintains the original structure, formatting, and meaning while incorporating shorter sentences, simpler language, and transition words for better readability. Each paragraph flows naturally into the next, creating a cohesive narrative about this controversial AI development.
This stance seems puzzling at first. The senator, known for opposing facial recognition by government agencies, now supports protective measures for law enforcement. In fact, while she fights against commercial and public use of these technologies, she backs their government use, as Politico points out.
Currently, no federal law clearly bans the use of public AI tools for identification. Thus, the ICE List project highlights a worrying gap in our laws.
Groups like the International Biometrics + Identity Association have created ethical guidelines for using biometrics and facial recognition. However, these rules don’t apply to Skinner’s case. After all, he’s not selling technology but running an informal activist project.
Skinner won’t share his system’s technical details. Still, he admits his tool only gives a rough “guess” of faces from ICE operation videos. Next, these images go through reverse searches on public platforms like PimEyes to match rebuilt faces with real people. Notably, federal police already use this method.
According to Politico, a 2019 study by the Georgetown Law Center on Privacy and Technology showed something interesting. Some police forces were changing photos or using sketches to identify suspects through facial recognition.
Photo: Joe Raedle/Getty Images.
Experts warn about this method’s reliability
Jake Laperruque from the Center for Democracy and Technology raises a key point. He says facial recognition becomes “highly unreliable” when it uses generated images instead of real faces.
Even Skinner admits his system’s flaws. In fact, it gets things wrong up to 60% of the time. Still, he ensures volunteers check each match before posting anything online. Also, he keeps sensitive details like addresses or personal numbers off his lists. “While I don’t support mob justice,” he states, “I believe in public shame and accountability.”
The rewritten version maintains shorter sentences and simpler language while keeping the original structure and meaning. It adds transition words like “currently,” “still,” “thus,” and “notably” to improve flow. Each paragraph naturally connects to the next, creating a clear narrative about this controversial AI tool and its implications.
This distinction, he argues, sets his project apart from aggressive doxxing. However, critics point out that even sharing just a name could put an agent and their family at risk.
An Open Political Battle
The ICE List project debate shows a deep political split. On one side, some Democratic lawmakers want ICE agents to clearly identify themselves. They believe this promotes openness and accountability. Meanwhile, Republicans push to make any identity exposure illegal, citing national security concerns.
So far, no federal law has passed on either side. Moreover, the lack of clear rules makes the situation more tense. At the same time, in the United States, it remains very easy to buy personal data if you have someone’s name.
“Anyone who wants to remove their information from the internet should have a fair chance to do so,” notes a data privacy expert. This simple statement highlights the complex balance between transparency and privacy in today’s digital age.
