The Download: investigating deaths at the US border’s “virtual wall”

The US government has invested billions of dollars in a high-tech system along its border with Mexico, often called the 'virtual wall'. This system includes nearly 600 surveillance towers equipped with cameras and artificial intelligence (AI) designed to automatically detect and track people crossing the border. The towers are part of a broader effort to monitor and secure the border without relying solely on physical barriers. Despite this investment, the system has repeatedly failed to prevent deaths, as shown in a recent investigation by MIT Technology Review. The investigation found that at least 138 cases since 2022 involved migrants whose remains were discovered within the range of a nearby surveillance tower, meaning they were likely visible to the system when they died.

In April 2024, José Morales Bernal, a 32-year-old man, crossed the US-Mexico border and was within range of three surveillance towers. These towers use AI to detect and track movement, and Border Patrol agents are trained to respond to alerts from the system. However, Morales was not apprehended, and he died just 360 feet (about 110 meters) from the closest tower. Local landfill workers, not Border Patrol agents, were the first to find his body. An autopsy later determined that Morales died from 'environmental exposure', meaning his death was likely caused by extreme heat, dehydration, or other harsh conditions at the border. This case highlights the system's failure to save lives despite its advanced technology.

A 15-month investigation by MIT Technology Review mapped nearly 4,000 locations where human remains were found along the US-Mexico border and compared these locations to the positions of the 600 surveillance towers. The investigation revealed that many deaths occurred in plain sight of the government's own cameras. The analysis showed that more than half of the people whose remains were found near towers were likely within the towers' field of view when they died. This suggests that the surveillance system is not only ineffective but also that its failures have a direct human cost. The investigation also found systemic flaws, including broken towers, AI algorithms that failed to detect people, and agents who did not respond to alerts.

Despite the clear failures of the virtual wall, the US government is planning to spend another $1 billion to triple the size of the surveillance system. Customs and Border Protection (CBP), the agency responsible for border security, has acknowledged some of the system's flaws, such as broken towers and non-functional AI. However, the investigation suggests that these issues are not isolated incidents but part of a larger pattern of systemic failure. The CBP has been criticized for not adequately maintaining the towers or ensuring that agents respond to alerts in a timely manner. The investigation proposes four ideas for how CBP could address these failures, but it remains unclear whether these changes will be implemented.

The failures of the virtual wall have had a devastating human impact. The investigation uncovered stories of migrants like Gómez Hernández, whose remains were found within the range of a surveillance tower but were not detected by the system. These failures are part of a broader humanitarian crisis at the US-Mexico border, where extreme conditions and lack of access to medical help have led to numerous deaths. The investigation is part of a series called 'Dying on Camera', which explores the failures of border surveillance technology and the human cost of these failures. The series aims to shed light on the shortcomings of the system and the lives lost as a result.

The surveillance towers along the US-Mexico border use AI to automatically detect and track people crossing the border. AI, or artificial intelligence, refers to computer systems designed to perform tasks that typically require human intelligence, such as recognizing patterns or making decisions. In this context, the AI is programmed to identify movement and alert Border Patrol agents. However, the investigation found that the AI systems often failed to detect people, either because the towers were broken or because the algorithms were not functioning correctly. This highlights the limitations of relying on AI for critical tasks, especially in high-stakes environments like border security.