US Military Explores Humanoid Robots for Potential Battlefield Roles
Companies are developing AI-driven robots for military applications, with testing underway, raising questions about the future of warfare and ethical considerations.
The U.S. military is actively experimenting with humanoid robots, exploring their potential integration into various roles, including those on the battlefield. Companies like Foundation, which develops AI systems for robotics, are conducting pilot programs with the U.S. military, alongside testing with Ukrainian forces. Foundation's work involves robots like 'Phantom,' directed by an AI system named Cortex, which is designed to perform tasks such as moving supplies or mapping buildings.
Sankaet Pathak, a co-founder of Foundation, argues that China's pursuit of this technology necessitates a response from the West. He envisions a future with hundreds of thousands of AI-driven humanoid robots forming ground forces, drawing a parallel to the increasing use of autonomous drones in the air. Pathak believes such a force could serve as a significant deterrent to conflict.
The U.S. military pilot program with Foundation focuses on robots handling, rather than firing, weapons. However, weaponization is part of the testing in Ukraine. This exploration aligns with broader military interest in robotics, exemplified by a U.S. Army contest seeking humanoids capable of supporting soldiers in diverse tasks.
Dean Fankhauser of Robozaps, a humanoid robotics advisory firm, stated that it is "completely inevitable" for companies to see business opportunities in weaponizing this technology. He noted that simpler robots, such as drones and ground systems, are already employed to carry explosives and missiles, with battlefield use particularly visible in Ukraine. Some firms are also working on weaponizing quadruped robots, though they have not been widely deployed in active warfare.
However, not all companies in the robotics sector support weaponization. Some legged robot companies have taken a stance against it, citing risks of harm and ethical concerns. Pathak, conversely, believes it is dangerous that more companies are not pursuing this path, arguing that humanoid robots are practical because the human-built world is already designed for human-like interaction and tool use.
Pathak emphasizes the importance of human oversight, stating that humans should be "in the loop" to approve any use of lethal force. He does, however, allow for exceptions where autonomous firing might be necessary to prevent catastrophic outcomes, and foresees scenarios where human authorization is less critical.
A primary challenge in developing these robots, shared across the industry, is creating artificial intelligence capable of functioning effectively in unpredictable real-world environments. Phantom's AI system, Cortex, uses a combination of a "reasoning model" trained on specific task examples and a broader "world model" trained on diverse data, including internet videos and the robot's physical interactions, to navigate and execute actions safely.
Foundation has secured $24 million in research contracts for its work. The company gained public attention earlier this year when Eric Trump, son of the former U.S. President, became an investor and advisor. Pathak's involvement also follows the bankruptcy filing of his previous financial services firm, Synapse, in 2024.