A year-long investigation reveals that billions invested in AI-powered border surveillance have not stopped hundreds of deaths, exposing fatal gaps in detection and agent response along the U.S.-Mexico border.

Introduction

The United States has invested billions in a high-tech surveillance network along its southern border, aiming to curb illegal crossings and save lives. However, a year-long investigation found that hundreds of deaths have occurred within view of these very systems, casting doubt on whether the technology is achieving its intended purpose.

What Happened

Since the 1990s, the U.S. government has deployed a layered surveillance network spanning nearly 2,000 miles along the southern border. The system includes earlier Remote Video Surveillance Systems, autonomous towers from Anduril, and integrated fixed towers from Elbit Systems. Despite their advertised ranges, the investigation identified over 1,050 deaths recorded within range of these towers between 2015 and early 2026. One prominent case involved José Morales Bernal, who died just 360 feet from the nearest Anduril tower in southern New Mexico. In another, a family of three crossing near Del Rio; the mother witnessed her daughter vanish into the river while cameras watched nearby, yet help arrived too late. Topographical analysis revealed that many towers see as little as 10% of their advertised surveillance area, and deaths still occurred outside documented blind spots.

Why This Matters

The investigation exposes a critical mismatch between official claims and on-the-ground reality. Government officials have long asserted the surveillance network improves detection and saves lives, but the findings show repeated failures: towers with clear views of death sites, agents unaware of people in distress, and no formal process to determine whether surveillance should have detected victims. With billions more allocated for thousands of additional towers through 2034, the report warns that without structural reforms, the virtual wall may continue to record fatalities rather than prevent them. Critics highlight the program's lack of transparency, independent oversight, and accountability mechanisms.

Key Takeaways

  • More than 1,050 people died within range of border surveillance towers between 2015 and early 2026, spanning both legacy and AI-powered systems.
  • Topographical analysis shows many towers observe as little as 10% of their advertised area, yet fatalities still occur outside recognized blind spots.
  • Over 110 deaths have been recorded within range of modern Anduril autonomous towers since 2021, despite claims of automated detection and tracking.
  • At least 350 deaths are linked to Elbit's integrated fixed towers, with some towers recording fatalities even when estimated to have clear line-of-sight to the location.
  • CBP has not conducted its own comprehensive analysis of deaths near its towers, and incident investigations into whether surveillance should have detected victims are rarely performed.
  • Billions in new funding are earmarked for 1,497 additional towers by 2034, even as the investigation highlights persistent gaps in detection, response coordination, and oversight.

Conclusion

As surveillance technology expands under growing federal investment, the investigation makes clear that more cameras and AI do not equate to fewer deaths. The human cost of the virtual wall's failures is counted in lives, and without greater transparency, independent evaluation, and reformed response protocols, the pattern is likely to repeat. The central question remains: when technology watches but no one arrives in time, who is truly responsible?