Analysts Warn of Cybersecurity Risks in Humanoid Robots

Analysts warn that humanoid robots bring a new layer of cybersecurity risk, with experts noting that securing these systems is exceptionally difficult because a robot is a system of systems—a network of networks including sensors, actuators, and a computation suite that must operate with millisecond speed.

The piece highlights that vendors struggle to implement robust authentication and encryption without hampering performance, and that many companies are still at early cybersecurity stages, sometimes relying on basic access controls rather than integrated protections.

It cites real-world concerns around Unitree, a Chinese vendor, where researchers have demonstrated how to root the robots and compromise devices within Bluetooth range, while also showing that internet-connected units transmit system data to servers without user consent.

The article also points to ongoing cyberespionage activity in the robotics sector, noting three or four long campaigns involving Russia’s DcRAT and AsyncRAT since the fall of 2024, and mentions more than 5,000 Chinese patents referencing the term “humanoid” in the past five years, according to Recorded Future.

According to Xinhua via Alamy Stock Photo, the report calls for advances such as a Secure Robot Operating System (SROS) built on ROS and a move toward zero-trust architectures to raise the bar for humanoid security.