RaonSecure Selected for Government Support Project with 'Homomorphic Encryption-Based AI Facial Recognition Technology'

IT security and authentication platform company RaonSecure (CEO Lee Soon-hyung, Lee Jung-ah) announced on the 26th that its self-developed “homomorphic encryption-based artificial intelligence (AI) facial recognition technology” has been selected for the “2024 AI Security Prototype Development Support Project” announced by the Ministry of Science and ICT (hereinafter ‘MSIT’) and the Korea Internet & Security Agency (KISA).

As the importance of strengthening response capabilities against cyber security threats grows, MSIT and KISA have been conducting this project annually since 2021 with the aim of supporting product development of promising domestic AI security companies and promoting the AI security industry.

RaonSecure has been jointly researching homomorphic encryption-based biometric information protection technology since last year with Cryptolab, led by Professor Cheon Jeong-hee of Seoul National University, in response to the increasing recognition of facial information as a major biometric and the growing importance of security in data processing.

RaonSecure plans to actively utilize the research and development know-how accumulated by its AI R&D team in carrying out this project. The company has been operating an AI R&D team composed of researchers with high expertise in AI technology, aiming to accumulate sustainable AI security technologies.

The AI R&D team is conducting research and development projects including image classification models that detect deepfakes, models that can detect AI-generated content, malicious and threat behavior classification models that detect malware, smishing, phishing, etc., and language models such as domain-specific large language models (LLMs) specialized for each company and enterprise LLMs.

In addition to the homomorphic encryption-based AI facial recognition technology development project selected for the KISA 2024 AI Security Prototype Development Support Project, RaonSecure’s AI R&D team plans to incorporate its self-developed “deepfake detection facial recognition technology” and “AI-generated content detection technology” into its existing consumer mobile security applications and unveil them in the second half of this year.

RaonSecure and Cryptolab signed a memorandum of understanding (MOU) in December last year to commercialize post-quantum cryptography (PQC) and homomorphic encryption (HE) technologies, establishing a close cooperative relationship through joint research and development aimed at cryptographic technology development. The two companies expect that through this project, they will contribute to the activation of the industry by presenting new security standards based on homomorphic encryption in the biometric authentication market, including establishing a personal facial information protection system using homomorphic encryption technology.

Lee Soon-hyung, CEO of RaonSecure, said, “As generative AI is rapidly spreading in the fields of security and identity authentication, we have formed an AI R&D team to research and develop related technologies and are achieving step-by-step results. Through facial recognition security technology applying generative AI and homomorphic encryption technology, we will further enhance the safety of identity authentication methods using biometric authentication.”