Aim Intelligence, LG Electronics Jointly Develop 'Physical AI Safety Layer'

A new horizon for robot safety is opening. AI security specialist Aim Intelligence is partnering with LG Electronics and Openmind, an innovative U.S. robot OS company, to jointly develop a ‘Physical AI Safety Layer’. This collaboration is expected to become a next-generation solution that completely protects 'Physical AI' operating in physical spaces from malfunctions or external attacks. Currently, safety standards for Physical AI remain at a theoretical level, such as Asimov's 'Three Laws of Robotics,' and have been criticized for their vulnerability to complex malfunctions and sophisticated external attacks in real-world environments. Furthermore, recent research has shown that attacks capable of disrupting AI and manipulating robot commands using ultrasonic signals imperceptible to humans or tiny image patches on clothing or helmets are becoming a reality, making the urgent establishment of a more sophisticated and robust safety net imperative. The new safety layer to be jointly developed by the three companies offers a fundamental solution to these threats. It will become a next-generation safety standard that proactively blocks robot judgment errors, or 'hallucinations,' and effectively defends against external intrusions and attacks. Specifically, through advanced technology that comprehensively analyzes visual, auditory, and sensor data, it will ensure robots always operate safely and reliably, protected from covert and unpredictable attacks. In this collaboration, each company will focus its core competencies. Openmind, which provides the universal robot OS ‘OM1’, will lay the foundation for this new safety layer. LG Electronics will be responsible for hardware development and real-world validation to verify the technology's practical applicability. Aim Intelligence will deeply integrate AI security technology into the robot operating system, designing a robust framework to ensure all robot actions are performed safely under human control. Ha-eon Park, CTO of Aim Intelligence, stated his ambition, saying, "Our goal is to build a system that goes beyond simply preventing malfunctions to completely blocking malicious external attacks. This technology is applicable across industries and will set a new standard for reliable robot operation even in autonomous environments." Meanwhile, Aim Intelligence, specializing in AI risk detection and blocking technology, has been recognized for its technological prowess at leading international conferences such as ACL and ICML. It is currently leading the industrialization of AI safety technology through collaborations with major domestic and international companies including KT, LG, BMW, Meta, KB Kookmin Card, and Korea Credit Information Services.