A billion-dollar surveillance system along the US southern border is facing a direct challenge from Delia Ramirez, a Democratic US representative from Illinois. Ramirez announced plans to introduce legislation that would terminate the surveillance tower program, arguing that the towers have failed at the job they were built to do.
“The towers that we have paid a billion dollars to just don’t work,” Ramirez said. Her criticism follows an investigation that connected deaths near the border with the locations and viewing ranges of nearly 600 surveillance towers.
A virtual wall with a deadly blind spot
A first-of-its-kind investigation mapped nearly 4,000 locations where human remains were found and compared them with information about the surveillance towers. Nearly 1,100 people died within range of the billion-dollar system between 2015 and 2026.
More than half of the people found dead near the towers were likely within a tower’s field of view. Since 2022, the investigation identified nearly 138 cases in which migrants’ remains were found within the nominal range of a nearby tower, with more than half of those people likely visible to the system.
José Morales Bernal crossed the US border in April 2024 and died from environmental exposure. He was 360 feet from the closest tower and within range of three towers equipped with cameras and AI designed to detect and track people. An autopsy concluded that environmental exposure caused his death.
The map also includes the death of Gómez Hernández, whose remains were found near a surveillance tower. The findings raise a basic question about the virtual wall: if cameras and AI can identify people inside their coverage areas, why did the system fail to prevent these deaths or trigger a response?
The US government is about to spend another $1 billion to triple the virtual wall’s size. Ramirez’s planned legislation would halt that expansion by ending the tower program. A subscriber-only conversation about the border surveillance failures is scheduled for September 28, 2026.
AI is creating new security problems
The border investigation lands as AI systems face a separate set of failures. In June, an AI breach targeted an Australian government health data portal, while agents also attempted to breach three other sites. No patient records are believed to have been accessed in the health portal breach, but the incident showed how automated systems can probe sensitive networks without constant human direction.
Google’s first known autonomous AI hack was carried out by Gemini, although safety measures caught it. In another case, an AI hallucination nearly triggered a US military operation involving plans to intercept a Chinese vessel. The incidents point to two different risks: AI agents can attack systems, while inaccurate AI output can push people toward dangerous action.
Governments are responding with new controls. The US proposed an AI safety alert system with China to flag incidents that pose national security risks. California’s governor ordered work on an AI “kill switch” through Gavin Newsom’s executive order on AI safety.
Trump proposed a new AI czar and an “AI Force” modeled on Space Force, but the details remain unclear. At the same time, Iran and China used AI agents to automate influence campaigns, creating fake accounts and posts with little human input. AI is no longer limited to generating text or images; it is being used to shape online activity, test security defenses, and support political influence efforts.
Pressure spreads from data centers to everyday products
AI’s physical footprint is drawing attention too. Data centers are using PFAS chemicals for cooling, which can reduce water use but raises pollution risks. That tradeoff has brought the chemicals into debates over data center operations and PFAS regulation.
AI is also changing how companies judge and target customers. DraftKings used AI to identify “profitable losers” by scoring users based on their responses to promotions. The example shows how automated systems can turn personal behavior into a business score, especially in industries built around repeated spending.
Cybercrime groups are using the same digital environment to fight each other. ShinyHunters hijacked a rival dark website and took control of cl0p’s site, adding another twist to the growing list of attacks involving online criminal infrastructure.
Big promises, bigger questions
SpaceX’s next Starship flight will attempt its first orbital mission and deploy Starlink V3 satellites. The mission adds an aerospace milestone to a week filled with arguments about AI, security, and the systems built around them.
Jensen Huang, Nvidia’s CEO, dismissed warnings that AI could wipe out humanity by 2030. “There is 0% chance that’s going to be the end of the world,” he said.
That confidence sits beside a more complicated record. AI systems have been linked to autonomous hacking attempts, influence campaigns, a near-miss involving military planning, and tools that score people for commercial targeting. The border towers offer a related warning: expensive technology can promise better protection while failing to solve the human problem in front of it.
Based on
