AI border surveillance fails to prevent migrant deaths

Death within range of surveillance

MIT Technology Review reports that Graciela Gómez Hernández, a 30-year-old migrant, died in September 2025 while attempting to cross the U.S.-Mexico border near San Diego. Her death occurred within one mile of a surveillance tower operated by General Dynamics, part of a network of AI-equipped border monitoring systems.

A family’s loss and the journey

Graciela sent voice messages to her family during her journey, including a final message stating she could no longer continue. Her sister, Irma, confirmed the details to MIT Technology Review. According to the report, Graciela had paid smugglers 20,000 pesos (approximately $1,150) to help her cross the border. After her death, her family paid over $4,000 to repatriate her remains to Mexico.

AI surveillance: billions spent, lives still lost

The surveillance towers deployed along the U.S.-Mexico border use artificial intelligence to distinguish between people, animals, and other objects. These systems, developed by companies like General Dynamics and Anduril, are part of a broader effort to create a "virtual wall" along the border. The U.S. government has invested over a billion dollars in this technology, with plans to spend an additional $1.4 billion to deploy 175 more towers.

Despite these investments, the technology has not prevented hundreds of deaths in recent years. According to MIT Technology Review, at least 138 migrants have died within range of surveillance cameras along the California border since 2022. More than half of these remains were found in locations where cameras would have had a clear view.

Systemic failures and unintended consequences

A 2024 report found that 30% of older surveillance towers were non-functional, rendering them little more than "million-dollar scarecrows." This raises questions about maintenance, oversight, and the effectiveness of relying on private contractors for border security infrastructure.

Migrants often take longer and more dangerous routes to avoid detection, increasing their risk of death from exposure or dehydration. Cesar Ortigoza, leader of the search group Armadillos, told MIT Technology Review that smugglers frequently push migrants into treacherous terrain, knowing they will be harder to detect.

Accountability and the human cost

The U.S. Border Patrol did not respond to questions about whether the tower near Graciela’s death detected her or why no action was taken. This lack of transparency is part of a broader pattern of accountability gaps in border surveillance operations.

Anduril, one of the companies involved in deploying these towers, claims its technology "directly contributed to saving lives and stopping illicit drugs." However, evidence suggests that the human cost of migration continues to rise despite these systems.

Human rights groups, including No More Deaths and Border Angels, argue that the focus on surveillance technology distracts from the need for humane immigration policies. Bryce Peterson, a researcher and volunteer with No More Deaths, said the technology creates a "cat-and-mouse dynamic" that pushes migrants into more dangerous situations.

The origins and expansion of border surveillance

The U.S. Customs and Border Protection (CBP) hasdeployed hundreds of surveillance towers along the southwestern border over the past decade. These towers, equipped with cameras, sensors, and AI, are designed to detect and track border crossers in real time.

The technology was initially tested in Afghanistan and other conflict zones before being adapted for border security. While it has successfully identified crossings, its record in preventing deaths remains questionable.

Critics, including the Electronic Frontier Foundation, argue that the expansion of surveillance infrastructure raises privacy concerns and disproportionately targets vulnerable populations. Dave Maass, the organization’s director of investigations, has called for greater transparency in how these systems are used.

What this means

LazyFounders analysis — our interpretation, not reported fact.

This story underscores a critical gap between technology and humanity in U.S. border policy. For founders and operators, it’s a stark reminder that deploying advanced systems—whether AI, surveillance, or automation—does not guarantee outcomes, especially in high-stakes environments. The failures here are not just technical but systemic: a lack of accountability, oversight, and ethical considerations in how these tools are deployed.

For startups working in govtech, defense, or AI, the lesson is clear: scalability and cost efficiency cannot come at the expense of reliability or moral responsibility. The border surveillance debacle also highlights the risks of over-reliance on private contractors, where profit motives may not align with public safety goals. Operators should ask: Are we building tools that genuinely solve problems, or are we creating expensive placebos that fail when it matters most?

Key takeaways

• Graciela Gómez Hernández died in September 2025 near a U.S. border surveillance tower, which may have detected her but did not prevent her death.
• Surveillance towers using AI have not stopped hundreds of migrant deaths along the U.S.-Mexico border in recent years.
• At least 138 migrants have died within range of border surveillance cameras in California since 2022, raising questions about their efficacy.
• The U.S. government plans to spend $1.4 billion on 175 additional surveillance towers, despite known failures in existing infrastructure.
• Migrants often take riskier routes to avoid surveillance, increasing the likelihood of death from exposure or dehydration.
• Border Patrol did not respond to questions about whether the tower near Graciela’s death detected her or why no action was taken.

Related on LazyFounders

• Anduril
• MIT Technology Review
• More policy news
• The US Border’s Virtual Wall: Billions Spent, Thousands Dead, and AI Failures Exposed