4 Ways to Address the Failures We Found Along the US Border's "Virtual Wall"
The US has spent billions on surveillance towers, but hundreds of people have died near them. Here's how to strengthen federal oversight.
By James O'Donnell | September 21, 2026
Introduction
Along the US–Mexico border, a network of surveillance towers—often called the "virtual wall"—has been deployed by Customs and Border Protection (CBP) at a cost of billions of dollars. These towers, equipped with advanced cameras, radar, and AI-powered analytics, are designed to detect and track movement across remote terrain. Yet an investigation has found that hundreds of people have died in areas monitored by these towers, raising serious questions about their effectiveness and the adequacy of oversight.
As of 2026, the virtual wall continues to expand, with CBP requesting additional funding for next-generation AI surveillance systems. But the human toll near these installations demands urgent reform. Below are four concrete ways to address the failures identified along the border's virtual wall.
1. Mandate Independent, Real-Time Auditing of Surveillance Data
One of the most glaring gaps is the lack of independent oversight of how surveillance data is collected, stored, and used. CBP's towers generate vast amounts of video and sensor data, often processed by proprietary AI algorithms. Without external audits, there is no way to verify whether these systems are accurately detecting distress signals, such as migrants in peril, or whether they are disproportionately flagging innocent activity.
Recommendation: Congress should require CBP to submit to annual, independent audits conducted by third-party technologists and civil rights experts. These audits must have access to raw data and algorithm logs, and their findings should be made public—with redactions only for genuine security concerns.
2. Integrate Humanitarian Response Protocols into Surveillance Systems
Surveillance towers are optimized for detection, not rescue. When the AI identifies a person crossing the border, the system often alerts Border Patrol agents, but there is no standardized protocol for triggering a humanitarian response—such as dispatching emergency medical services or notifying local search-and-rescue teams. In the harsh desert environment, minutes matter.
Recommendation: CBP should integrate automated humanitarian alerts into the virtual wall's architecture. When a person is detected in distress—e.g., stationary for an extended period, moving erratically, or in extreme heat—the system should simultaneously notify both Border Patrol and nearby emergency responders. This dual-notification approach has been piloted in some sectors but must be mandated nationwide.
3. Establish Transparent Reporting on Deaths and Near-Death Incidents
Currently, data on deaths near surveillance towers is fragmented and inconsistently reported. Advocacy groups and journalists have had to piece together information from local coroners, CBP press releases, and nonprofit organizations. This lack of transparency makes it impossible to assess the true impact of the virtual wall or hold agencies accountable.
Recommendation: CBP should publish a quarterly public report detailing all deaths and near-death incidents within surveillance zones, including the location, time, cause, and whether the tower's AI system detected the individual before their death. This data should be standardized and cross-referenced with autopsy reports. In 2026, with AI systems under increasing scrutiny, such transparency is not just ethical—it is essential for public trust.
4. Overhaul the Procurement Process to Prioritize Human Rights Impact
The companies that build and maintain the virtual wall—such as Anduril, Elbit Systems, and General Dynamics—are profit-driven and often prioritize technical performance metrics over humanitarian outcomes. Contracts are awarded based on detection accuracy, uptime, and cost, with little to no consideration of how the technology affects vulnerable populations.
Recommendation: Congress should amend procurement rules to require human rights impact assessments before any new surveillance contract is awarded. These assessments should evaluate how the technology might contribute to deaths, injuries, or civil liberties violations, and should include input from local communities and human rights organizations. Companies with a history of complicity in abuses should be disqualified.
Conclusion
The virtual wall is not just a technological system—it is a policy choice with life-and-death consequences. The failures documented along the border are not inevitable; they are the result of inadequate oversight, misaligned incentives, and a lack of accountability. By implementing these four reforms, the US can begin to ensure that its border surveillance infrastructure serves both security and humanitarian goals. As AI becomes ever more pervasive in 2026, the stakes could not be higher.
This article is part of the "Dying on Camera" series, examining the human impact of border surveillance technology.