AI in Immigration Tech

At the end of 2023, after being laid off from my role at a healthcare analytics company, I began researching CBP One: a Customs and Border Protection app originally designed as a single access point for CBP services. I first looked into it as a UI/UX case study, focusing on how its facial recognition requirements shaped the experience of migrants using the app to seek legal entry at the southern border.

What I found and documented was a sprawling digital paper trail tracing the app’s evolution from its October 2020 launch to its cancellation on January 20, 2025. At the time of its demise, CBP One served as effectively the only route for undocumented immigrants to legally cross the U.S. southern border.

After Donald Trump canceled the app on Inauguration Day of his second term as president, it was resurrected as CBP Home—built on much of the same technology, but now aimed at tracking the “self-deportation” of immigrants as an alternative to forced removal by Immigration and Customs Enforcement (ICE).

The user base for both apps is staggering:

  • Six days before Trump’s second inauguration, CBP issued a press release announcing that “more than 936,500 individuals have successfully scheduled appointments [via the CBP One app] to present at ports of entry instead of risking their lives in the hands of smugglers.” (DHS Secretary Kristi Noem would later state that CBP One “allowed more than one million aliens to illegally enter the country.”)
  • A year later, the DHS Office of Public Affairs issued a press releasestating that there have already been “nearly 100,000 users of the CBP Home app.”

CBP One’s AI facial recognition has drawn criticism since it was first deployed to verify undocumented immigrants’ identities in early 2021. Today, that same class of technology fuels the federal deportation effort’s surveillance capabilities. These tools can be used to target anyone in America, immigrant or no, and are currently in use doing just that.

This page curates my ongoing research into how DHS uses AI to gather “personally identifiable information that belongs to individuals who are not suspected of any wrongdoing,” as U.S. lawmakers recently put it.

The posts gathered here document why they’re asking questions — and what the available evidence already tells us.