The Virtual Border Wall Watches Death Without Responding

The virtual wall is missing people. The United States has spent billions over the past 25 years building surveillance towers along its southern border, promising a technical answer to border security. Yet hundreds of people have died near those towers without being reached or apprehended, including people watched by newly installed AI-powered systems designed to spot them automatically.

A 15-month investigation mapped nearly 4,000 locations where human remains were found against information on nearly 600 surveillance towers. The findings show more than 1,050 people are estimated to have died near towers, though that figure is likely an undercount.

More than 10,000 people have died during crossings since 2000, according to Border Patrol estimates. A separate analysis estimated that more than half were likely within a tower’s field of view — a grim achievement for a system built to see.

Technology That Sees Without Acting

In April 2024, José Morales Bernal crossed the border into the United States and died of environmental exposure 360 feet from the closest surveillance tower. Deaths like Morales’s are startlingly common, and more than 18 people had died near surveillance towers in one small stretch of desert during 2024.

The towers do not need to miss every person to fail. They only need to detect people without producing a response, leaving the system with an impressive view and an unimpressive record of intervention.

Anduril is building surveillance towers, including AI-powered towers that automatically identify people. The government is about to spend another $1 billion to triple the virtual wall’s size by 2034, even as the number of towers remains unclear: the investigation found nearly 600, while the government states that about 800 are in operation.

The gaps are not limited to the hardware. Customs and Border Protection has never conducted a comprehensive audit of the towers’ effectiveness since their installation began in 2015, and as of August 2021, it had not produced a public analysis of whether the towers worked.

A Broken Record of Failure

CBP’s system for tracking migrant deaths is incomplete and broken, with no shared protocol for documenting them. Border Patrol established its internal system, BSITS, in 2007, but many cases are not tracked properly; in 2021, the Government Accountability Office found that many migrant deaths were missing from it.

In 2023, the GAO reported that Border Patrol had taken steps to improve data collection, but its estimates still undercounted deaths. CBP launched the Missing Migrant Program in 2017 to prevent deaths, yet the program lacks a plan to measure whether it works.

The agency has also failed to reliably track when surveillance technology helps with apprehensions. The GAO suggested that practice in 2014, then reported in 2017 that apprehension data was inaccurate and unreliable.

CBP launched a program in 2025 to improve data collection about events such as rescues and 911 calls. That is an improvement, but it does not repair the older problem: the agency still lacks a dependable way to connect deaths with the surveillance system watching the border.

Surveillance footage from towers is typically deleted after 30 days unless an internal review begins. Those reviews usually happen only for deaths in custody, meaning footage that might show a failure to detect or respond can disappear before anyone examines it.

CBP should automatically check every migrant death against its internal database of towers. That basic comparison could show whether a person was inside a tower’s field of view, whether the system detected them, and whether anyone responded.

Sarah Silva, a New Mexico state representative, put the situation more plainly: “If I were the federal government, I would do an audit, because this feels like it’s not working.” Spending another $1 billion before conducting that audit would turn a surveillance gap into policy.

Stay connected via Google News

Follow us for the latest AI News