MIT Review Investigation Links Border Surveillance Towers to Migrant Deaths in Remote Terrain
An MIT Technology Review probe found that people crossed areas covered by AI-enabled border towers undetected and later died nearby, with remains sometimes undiscovered for long periods.
MIT Technology Review published an investigation today into the network of surveillance towers the U.S. government has built along the U.S.-Mexico border, finding that migrants have moved through areas monitored by the AI-enabled systems without being detected and later died near those same sites. In several cases, the outlet reports, the bodies went unnoticed for extended periods despite the towers' sensors operating nearby. The findings raise questions about whether the surveillance architecture, often described as a "virtual wall," is delivering on its stated purpose of spotting crossings and directing responders.
The towers are part of a program that Customs and Border Protection has expanded over the past several years, using a mix of cameras, radar, and machine-learning software to flag movement across remote stretches of desert, mountain, and ranch land. Contractors including Anduril Industries have supplied autonomous surveillance platforms that the agency says can detect and classify activity without a human watching every feed. The pitch has been that fewer agents are needed in punishing terrain and that help can reach people faster when something goes wrong.
What the MIT Technology Review reporting suggests is a gap between detection and response. A tower can register a heat signature or a shape moving through brush, but that alert still has to travel through a chain of monitoring, verification, and dispatch before anyone physically arrives. In the cases the outlet examined, that chain either never produced a response or produced one too late, and the terrain's remoteness meant no one else was likely to pass by.
The investigation arrives as border technology spending faces heightened scrutiny in Washington. Lawmakers have pushed for more autonomous systems along the southwest border, while human-rights groups and some local officials in border counties have argued that surveillance expansion outpaces the medical, search-and-rescue, and humanitarian resources needed to act on what the sensors find. Deaths in border regions have remained a persistent concern, with remains frequently recovered in areas far from roads and water.
The report also underscores a measurement problem: it is difficult to assess how well the towers perform because detection data and response outcomes are not routinely published in a form outside researchers can audit. Without that, claims about the system's effectiveness rest largely on agency statements and contractor figures, and the human cost of any gaps is recorded in county morgues rather than performance dashboards.
What to watch next is whether CBP or its contractors release detection and response data, and whether Congress attaches reporting requirements to any new surveillance funding. The deeper question the investigation leaves open is what a tower is actually for: flagging movement, or getting a person found alive.