● Covered by 1 source · 1 reportMedium impact1 neutral
Manipulated Inputs Can Bypass Robot Safety Systems, VicOne Research Shows
Key points
- Robots rely on inputs like distance, position, or stop signals for safe operation.
- Manipulated visual or audio inputs can alter robot behavior and task execution.
- VicOne LAB R7 tests showed robots misinterpreting text and inaudible audio as instructions.
- Researchers bypassed a robot's safe-control setting by injecting a ROS 2/DDS message.
The Challenge of Untrusted Robot Inputs
Robots make decisions based on interpreting their surroundings through various inputs, such as distance readings, position estimates, and stop signals. If these inputs are inaccurate due to faults or deliberate manipulation, a robot may respond as designed but in an unsafe manner. For example, a mobile robot might perceive a hallway as clear when someone is present, or a collaborative arm might misjudge a worker's proximity.
How Manipulated Inputs Alter Robot Behavior
Research indicates that information placed in a robot's environment can influence its actions. Studies on vision-language-action (VLA) models have shown that adversarial images can reduce task success or cause models to ignore instructions. VicOne LAB R7 conducted tests where text on a poster was interpreted as an instruction by a robot-dog using Gemma 4 E4B, altering its movement. In a hospital-service-robot simulation using Nemotron on NVIDIA Jetson AGX Orin, crafted audio inaudible to humans changed the robot's simulated behavior, despite assigned tasks remaining the same.
Vulnerabilities in Safety-Related Decisions
Even independent protective systems designed to prevent dangerous actions rely on accurate information, including readings, estimates, and messages. If these inputs are manipulated, the safety system itself can be compromised. During a robotics bug bounty event, VicOne LAB R7 researchers injected a ROS 2/DDS message into a robot that was expected to remain still under a safe-control setting, causing the robot to move.
Implications for Robot Safety
These findings highlight a critical security concern for robotics. The ability to manipulate sensor data or command inputs means that robots, even those with built-in safety protocols, could be tricked into performing unsafe actions. This necessitates a deeper focus on securing the integrity of robot input channels and the data processing within safety-critical systems.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
Reporting from
VicOne LAB R7 research demonstrates that robots can misinterpret their environment or receive false instructions through manipulated inputs, leading to unintended and potentially unsafe behaviors even when safety functions are active. This highlights a vulnerability where adversarial inputs can compromise robot safety systems that rely on accurate sensor data and commands.