TechTrendsLab
AI

The AI Virtual Wall Failed. Here's How to Fix It.

MIT Tech ReviewTuesday, September 22, 20263 min read
A tall surveillance tower with cameras overlooking a desert landscape near the US-Mexico border

The US-Mexico border's "virtual wall" of AI-powered surveillance towers was sold as a force multiplier. But a year-long investigation by MIT Technology Review and Times of San Diego reveals a darker reality: hundreds of people walked undetected through surveilled areas and later died nearby, their bodies undiscovered for weeks. With $1 billion slated to triple the tower network by 2034, the failures demand urgent answers.

A Surveillance Network That Doesn't See

The investigation found systemic flaws in the virtual wall: broken towers, algorithms that failed to detect people, and agents who didn't respond to alerts. In one small stretch of New Mexico desert alone, more than 18 people died near surveillance towers in 2024. Despite analyzing nearly 600 towers, the true number of deaths is likely higher because CBP doesn't publish a comprehensive map of its network. The government says 800 towers operate today, but without an internal audit, no one knows exactly where they are or how often they fail.

The Data Problem Nobody Wants to Fix

Tracking migrant deaths along the border is fundamentally broken. There's no national agency collecting and publishing all records, and Border Patrol's internal system, BSITS, only got national entry rules in 2021. The GAO found many cases still aren't tracked, especially deaths not discovered by Border Patrol. Meanwhile, the Missing Migrant Program launched in 2017 has no plan to measure whether it's actually reducing deaths. CBP also agreed in 2014 to log when surveillance tech led to apprehensions, but a 2017 GAO report found the data unreliable—agents were crediting towers that didn't exist in Texas.

What CBP Should Do Next

MIT Technology Review outlines four recommendations. First, conduct an internal audit to map every tower and quantify deaths nearby. Second, fix the broken death-tracking system so data is complete and shared. Third, require agents to check every migrant death against the tower database to see if surveillance detected them. Fourth, preserve footage from towers near deaths, which is typically deleted after 30 days. These steps would answer the question families and researchers keep asking: Did the cameras see them, and nobody came? Or did nobody see them at all?

Key Takeaways

  • Over 1,050 people died near AI surveillance towers, but the true toll is likely higher due to missing data.
  • CBP has never conducted an internal audit of the virtual wall's effectiveness despite having the capability since 2015.
  • The US plans to spend $1 billion to triple the virtual wall by 2034 without assessing its impact.
  • Agents aren't required to check if migrant deaths occurred near surveillance towers, leaving critical evidence unexamined.
  • Footage from towers is deleted after 30 days, often before families can request it.

Source: MIT Tech Review • 🇺🇸 San Francisco

Share:
#ai ethics#border surveillance#cbp#migrant deaths#virtual wall

Keep Reading

Related Articles