Along the U.S.-Mexico border, hundreds of AI-assisted surveillance towers now operate, capturing real-time images and data to monitor border activity, according to current and former officials familiar with the program. These towers, equipped with radar and high-resolution cameras, use artificial intelligence to identify and classify moving objects such as humans, animals, or vehicles, alerting Border Patrol agents to potential security concerns.
One such tower near Sunland Park, New Mexico, close to El Paso, sits on elevated terrain, allowing it to scan the border wall and surrounding city areas up to approximately two miles away. After detecting significant movement via radar, the tower’s camera zooms in to generate detailed images, which an AI model then analyzes to classify the target. Border Patrol agents reviewing footage at command centers use these alerts to determine appropriate responses.
Customs and Border Protection (CBP) oversees one of the largest federal visual surveillance programs currently in use. The agency emphasizes that AI technology enhances 24-hour monitoring capabilities with greater efficiency than human observers, who require rest and can be distracted. CBP Assistant Commissioner Hilton Beckham described the towers as “essential to border security,” stating that the technology aids agents in assessing risks and helps protect law enforcement personnel, migrants, and the public. She noted that CBP operates within guidelines aimed at respecting civil liberties and privacy interests.
The agency has the capacity to restrict access to certain camera views—a feature used to block sensitive areas from surveillance—but can also lift these restrictions as needed, though CBP has not publicly disclosed details regarding the frequency or locations of such measures. Surveillance footage collected through the towers may be shared with local, state, and federal law enforcement agencies, as well as foreign governments, according to CBP privacy guidelines. The agency acknowledges that people in border regions cannot opt out of being recorded and that recordings may include those participating in lawful, First Amendment-protected activities.
Since 2018, CBP has expanded its AI tower program to at least 350 units spanning remote and populated sectors along the southern border, from California to Texas. An analysis identified 209 of these towers—situated on federal land or leased properties—with an estimated combined surveillance reach encompassing approximately 180,000 U.S. residents and 850,000 residents in Mexico. The towers vary in height from roughly 18 to 120 feet and incorporate panoramic radar, powerful day and night cameras, onboard AI computing, and communication antennas. Importantly, CBP states that these towers do not employ facial recognition technology.
Despite the technological advances these towers represent, independent reviews have highlighted limitations. The Government Accountability Office has criticized the program for lacking a master plan outlining costs, timelines, and success metrics and noted gaps in policies for data security. Officials also acknowledge the towers’ reduced effectiveness in congested urban border areas due to higher volumes of people, vehicles, and construction activities.
Opinions among border communities differ regarding the surveillance towers. Some residents support the enhanced security amid increased migration during the Biden administration, while others express concern about potential infringements on civil liberties. Public understanding of the towers’ capabilities remains limited; many residents mistake them for cell towers or radar installations.
The majority of the towers were constructed by Anduril Industries, an AI defense firm with previous contracts during the Trump administration and continued work under the Biden administration. Anduril recently secured contracts valued at over $567 million and has been joined by three other companies authorized to build similar systems. Data gathered by the towers is stored on government servers, typically retained for 30 days unless needed for law enforcement or AI training, with contractors sometimes granted access under CBP authorization.
Critics, including civil liberties advocates, have labeled the program as expensive and ineffective, citing oversight reports that have repeatedly questioned its utility. The Electronic Frontier Foundation’s investigations director, Dave Maass, described the initiative as a recurring government expenditure marked by overstated benefits.
Supporters argue that when strategically deployed and combined with sufficient personnel resources, AI towers effectively deter unauthorized crossings in certain areas. Former Border Patrol official Rush Carter cited increased apprehensions near Valentine, Texas, after the installation of such a tower.
As border surveillance technology continues to evolve, CBP is working toward integrating data from various sources—including drones, license plate readers, trail cameras, and ground sensors—into a unified operational platform for greater situational awareness. Meanwhile, residents living near the border continue to adjust to the presence of these sophisticated monitoring systems, whose reach extends across communities on both sides of the Rio Grande.