East Security is collaborating with Naver Cloud to develop a domestically produced artificial intelligence (AI) foundation model specialized in cybersecurity. The project will leverage malware data, cyber threat intelligence (CTI), and breach analysis expertise accumulated through endpoint security operations to create an AI model applicable in real-world security monitoring and vulnerability analysis.
On September 7, East Security, a subsidiary of Eastsoft specializing in security, announced that the Naver Cloud Consortium, in which it participates, has been selected as the final executing agency for the 'Cybersecurity-Focused AI Foundation Model Development' project, overseen by the Ministry of Science and ICT and the National IT Industry Promotion Agency (NIPA).
This initiative is a government-led project aimed at responding to increasingly sophisticated cyber threats through the advancement of high-performance AI and securing a unique domestic security-focused AI model. The Naver Cloud Consortium, recognized for its superior technology and development strategy, was chosen as the final contractor.
East Security will provide extensive malware data and CTI expertise accumulated over many years in endpoint security to the Naver Cloud Consortium. This collaboration aims to implement an AI model that can be immediately utilized for security monitoring and vulnerability analysis in a top-tier 'full-stack' environment.
Recently, East Security has been leading the commercialization of domestic AI security technology by launching the 'Al-Yak Agent EDR' based on its proprietary foundation model 'K-Exa One' and signing a memorandum of understanding with a domestic AI semiconductor (NPU) company for the development of next-generation K-AI security solutions and technical cooperation.
With this project, East Security aims to maximize synergy by integrating domestic AI accelerators into its AI solutions, laying the groundwork for establishing a sovereign AI infrastructure specialized in domestic AI security.
In this project, East Security will not only utilize AI models but also provide high-quality cybersecurity big data accumulated from various security products and services, including the vaccine 'Al-Yak.' This will enhance the domestic AI security response capabilities, which have been relatively lacking, by combining endpoint threats, CTI, breach analysis, and analytical expertise.
Jung Sang-won, CEO of East Security, stated, "We will mobilize our proven capabilities in commercializing AI security services, as demonstrated by the launch of Al-Yak Agent EDR, along with our unique security data and analytical expertise to elevate national cybersecurity capabilities to a new level."
The Naver Cloud Consortium plans to develop two types of security-focused AI models based on HyperCLOVA X and Exa One, featuring a mixture of experts (MoE) structure with 700 billion parameters. Approximately 830TB of security and industrial data will be utilized for training, and the developed models will be tested in key industries such as power, finance, science and technology, telecommunications, semiconductors, defense, and aerospace. The project will run from this month until July of next year. The government will support the model development with 256 NVIDIA B200 GPUs, and the results will be made open-source to promote the domestic security ecosystem.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.