Raon Secure is spearheading a collaborative project to establish a security framework for the era of physical artificial intelligence (AI). The company plans to equip robots and AI agents with digital identities to verify their identity and authority while developing technologies to address potential security threats in industrial settings.
On October 15, Raon Secure announced that it has been selected as the lead organization for the 'Development of Multi-layered Security Technology Based on Physical AI for Collaborative Intelligent Systems,' a project supported by the Ministry of Science and ICT and the National IT Industry Promotion Agency (NIPA).
Physical AI refers to technology that allows AI to directly control real-world devices, such as robots, to perform physical tasks. Cyberattacks or incorrect AI judgments can lead to equipment malfunctions, production halts, and safety incidents. In industrial environments where multiple AIs and robots collaborate, a security system is essential to verify the identity, authority, commands, and judgments of each entity.
In this project, Raon Secure will develop technologies to systematically verify and respond to security threats arising from AI models, data, communication, devices, and operational environments.
Raon Secure will oversee the entire research and development process and focus on developing identity and authority management technologies for physical AI robots and agents. The company plans to utilize decentralized identity (DID) technology, which has been applied to national mobile identification, to provide unique digital identities to robot agents and establish a system for managing authority, delegation relationships, and activity histories.
This will enable real-time verification of qualifications and authority when robots operate machinery, controlling commands that exceed permitted limits at the execution stage. The DID technology, previously used for human identity verification, will be expanded to manage the identity and actions of machines.
Participants in this research and development project include Sensetone, Thillon, Disilo, IT Station, Sogang University, and Jeonbuk National University.
Sensetone will focus on access authentication using one-time authentication code (OTAC) technology and detecting and blocking abnormal commands, while Thillon will handle the detection of anomalies in isolated work environments. Disilo will work on data protection using homomorphic encryption (HE), and IT Station will conduct field demonstrations in manufacturing and robotics.
Sogang University will develop technology to verify the integrity of AI judgment results, and Jeonbuk National University will create technology to assess the stability of AI models in changing environments or during cyberattack scenarios. The research team plans to combine these efforts to implement a physical AI security platform that protects AI models and field devices.
Based on these technologies, the team aims to develop a physical AI security platform that integrates an 'intelligent testing module' for preemptively assessing AI model safety and a 'trust integrity module' for real-time protection of field devices. The scope of application is expected to expand from smart factories and autonomous manufacturing to logistics, healthcare, and defense.
Raon Secure has been involved in the government mobile identification project, applying DID-based identity verification technology. In September, the company unveiled 'One Access,' which extends this technology to manage digital identities, authority, and activity histories in the realms of AI and robotics.
Additionally, in May, Raon Secure signed a memorandum of understanding with the Korea Electronics Technology Institute (KETI) to enhance physical AI security technology. Both parties are jointly verifying the necessary security standards and specific application methods to ensure the stable implementation of agentic AI technology controlling physical systems like robots and drones in industrial settings.
Lee Jeong-a, CEO of Raon Secure, stated, "The success of the physical AI era depends not on how advanced it is, but on how trustworthy it is. We aim to extend the authentication technology verified through the national mobile identification to the realm of robots, establishing security standards that allow industries to confidently rely on AI and machines."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.