AJU PRESS ASIA FIRST NEWS AGENCY

Journalist

Kim Ki-eung
  • Five Firms Vie for South Korea’s $2 Billion GPU Project as Next-Gen Platform Looms
    Five Firms Vie for South Korea’s $2 Billion GPU Project as Next-Gen Platform Looms Competition is intensifying among major companies for the government’s 2.08 trillion won project to secure advanced graphics processing units, with the ability to build large clusters, whether to adopt Nvidia’s next-generation “Vera Rubin” platform, and how much of the GPUs bidders plan to use themselves emerging as key factors. Yonhap News reported on Saturday that the National IT Industry Promotion Agency (NIPA) plans to select operators next month for the “GPU procurement, buildout and operations support” project. Led by the Ministry of Science and ICT, the initiative aims to quickly secure large-scale GPUs through public-private cooperation and supply them to domestic industry, universities and research institutes. The project centers on using government funds to secure about 15,000 of the latest GPUs and choosing private cloud providers to build and run them reliably. Five companies submitted bids: Naver Cloud, KT Cloud, Samsung SDS, Coupang and Elice Group. Industry officials expect multiple operators could be chosen again, as in last year’s 1.4 trillion won program, which selected Naver Cloud, Kakao and NHN Cloud. A central issue this time is how the GPU volume will be allocated. Last year, based on proposed scale and build capability, NHN Cloud received about 8,000 units, Naver Cloud about 3,000 and Kakao about 2,400, with NHN Cloud winning the largest share after its large-cluster plan scored highly. Cluster size remains a major evaluation item. The government requires at least one cluster of 256 nodes, or 2,048 GPUs, or more, because larger clusters significantly improve processing speed and efficiency. Another new variable is whether bidders will adopt Nvidia’s “Vera Rubin” platform. The government plans to award extra points to proposals that include it. Naver Cloud and Elice Group are reported to be reviewing adoption, while some companies are said to have excluded it. Vera Rubin’s high power density and facility load requirements could limit where it can be deployed, raising concerns that adoption will depend heavily on each company’s data-center infrastructure. Strategies also differ, with some proposing to use existing data centers for large GPU clusters and others promoting new approaches such as modular data centers. How much of the GPUs companies plan to use themselves is also seen as decisive. The government will purchase and retain ownership of the GPUs, but participating companies may use the remainder after meeting public supply requirements. Because the government gives extra points to bidders offering a higher share for public supply, companies must balance profitability against evaluation scores. Industry sources said last year’s participants set self-use at about 20%, limiting profitability. “Once you factor in adopting a next-generation platform, the investment burden rises sharply,” an industry official said. “It’s not an environment where it’s easy to make aggressive large-scale proposals.”* This article has been translated by AI. 2026-04-26 11:27:09
  • Teacher Says AI Platform Turned Static Slides Into Student-Led Classroom Discussion
    Teacher Says AI Platform Turned Static Slides Into Student-Led Classroom Discussion South Korea introduced a requirement in the 2022 revised national curriculum for 34 class sessions of information education in elementary schools. But specific teaching methods and materials are still largely left to teachers. Oh Yuna, a teacher at Hongneung Elementary School in Seoul, said she has been filling that gap with a platform called Sooup AI ("Lesson AI") from startup Redbrick. Oh, who has received a minister of education commendation for work in strengthening teachers’ digital capabilities, said she has seen clear changes in how classes run. She spoke about those changes in an interview. ■ From static slides to discussion… "It takes one click" "Lesson AI" lets teachers upload existing PPT or PDF materials. The system analyzes them and automatically converts them into student-participation activities such as quizzes, discussion prompts and question-based tasks. Oh said it reduces preparation time and also helps teachers adjust lessons on the spot. "Most materials teachers have are PPTs or PDFs, and Lesson AI creates activity materials for students based on those lesson plans," she said. "Because the AI automatically builds activities like quizzes or discussions, the burden of making new materials drops a lot." During class, she said, it can function like an assistant. "When I’m teaching and want to reinforce a concept, Lesson AI generates that content right away," Oh said. "Being able to expand existing materials instantly is very useful." She also cited its ability to analyze individual learning data and provide immediate feedback, especially in subjects where student gaps are wide. Oh said the most visible change has been student participation. "A child who usually wouldn’t write an opinion in a textbook will actively type their thoughts in a discussion using digital devices," she said. "It seems speaking and expressing yourself in writing work differently for kids." She added that what matters is not just who speaks up, but "how much they are actually thinking." Using Lesson AI, she said, teachers can check multiple students’ thought processes in real time, drawing out a different kind of cognitive engagement than in traditional classes. She said students’ reactions have been positive, with comments like, "Let’s do this again," after class. "You can see kids trying to participate more actively through the tool," she said. Oh said the technology has also changed how teachers use their time. "Before, I had to spend a lot of time analyzing the curriculum and preparing materials, but now the AI takes over part of that process," she said. "That lets me use the time I gain to understand children’s characteristics and the classroom atmosphere more deeply." She said students, too, are beginning to set their own learning direction as real-time data helps them see their level more objectively. Addressing concerns from some parents about overreliance on devices, Oh said use is limited. "We don’t use devices for the entire class, only at necessary moments," she said. "Because activities are balanced — research, organizing opinions and presenting — situations where kids get overly absorbed in devices actually decrease." ■ Data resets each school year… "A condition for real personalized education" Oh said there are still limits in public education. "In AI education, accumulating data is very important, but right now the structure doesn’t carry data over when students move up a grade," she said. "It’s not easy because of privacy or system issues, but data has to accumulate for true personalized education to be possible." She said the most important factor is still teachers’ ability to design lessons. "No matter how many tools there are, if you don’t have the ability to use them for the children in your class, it’s hard to sustain," Oh said. "This is an opportunity to look back at the classroom and place technology appropriately where it’s needed."* This article has been translated by AI. 2026-04-23 15:10:34
  • Kakao Speeds Global Push With KakaoBank and KakaoPay, Linking AI and Payments
    Kakao Speeds Global Push With KakaoBank and KakaoPay, Linking AI and Payments Kakao Group is accelerating its global push by leaning on its financial affiliates as it moves to execute its next-generation finance strategy. Kakao said Wednesday that KakaoBank and KakaoPay are splitting roles overseas to strengthen the group’s financial competitiveness. The effort follows the “global fandom OS” strategy Jeong outlined earlier this year. The plan aims to expand Kakao’s global user base by combining agent artificial intelligence, entertainment intellectual property and Web3 technology. In finance, Kakao has positioned AI-driven services and the buildout of global payment infrastructure as key pillars. KakaoBank is exporting its digital finance model, focusing on Southeast Asia and Central Asia. In Indonesia, it entered the market by investing in the digital bank Superbank in partnership with Grab. In Thailand, it is working with a local partner to establish a virtual bank. More recently, it has teamed with Mongolia’s MCS Group to invest in the digital bank M Bank and to upgrade credit-scoring models, part of a broader push to spread an inclusive finance model globally. KakaoPay is expanding payment infrastructure overseas to speed its shift into a “global pay” platform. After rolling out QR payments, it introduced NFC payments to enable use at major merchants worldwide, and it is focusing on providing the same user experience abroad as in South Korea. It is also covering payments by foreign visitors in South Korea, strengthening a two-way payments ecosystem. KakaoPay recently joined the x402 Foundation, led by the Linux Foundation, to help build next-generation payment standards. It is also expanding points of contact with Circle, a global stablecoin company, as it explores ways to broaden a digital asset-based financial ecosystem. Kakao said its integrated approach across banking, payments and platforms is accelerating work on a new model that links AI, finance and payments. The company said that if “agentic finance,” in which AI takes part in financial decision-making, is combined with global payment infrastructure, changes to the existing structure of the financial industry could follow. 2026-04-23 10:21:06
  • Hancom to Unveil ‘Twin Agentic OS’ Aimed at 24/7 AI-Assisted Work
    Hancom to Unveil ‘Twin Agentic OS’ Aimed at 24/7 AI-Assisted Work Hancom, formally known as Hancom Inc., said it will introduce what it described as South Korea’s first “twin agentic OS,” designed to reflect a user’s work patterns and support around-the-clock productivity. The company said AI agents would be able to carry out tasks autonomously even after users leave work. Hancom on the 23rd shared companywide results of its AI transformation, or AX, and outlined its strategy at an event called “Hancom AX Day.” CEO Kim Yeon-su said the company plans to reveal the “twin agentic OS” in the first half of the year and commercialize it within the year. At the core of the system is an AI agent that mirrors a user’s work style — a “digital twin,” the company said. By learning an individual’s work methods and patterns, the agent can continue handling tasks when the user is away. Hancom said the system is built on its document-structuring technology accumulated over 36 years and its AI capabilities, and is intended to serve as an “intelligent control tower” that links various AI models with existing workplace systems. Kim said the company will move beyond a business centered on packaged software and shift into building AX execution systems that let customers directly experience productivity gains in the field. Hancom also said it will pursue organizational culture changes alongside technology adoption. It plans to run an “AX Champion” program to select and reward employees who deliver measurable workplace improvements, and to promote AX through performance-based evaluations rather than whether employees use AI tools. “AX is a change that reshapes the overall way we work,” Kim said, adding that the company aims to establish itself as a trusted AX partner through the twin agentic OS. * This article has been translated by AI. 2026-04-23 08:57:16
  • Naver to Disable Comments on Articles Flagged for Excess Malicious Posts
    Naver to Disable Comments on Articles Flagged for Excess Malicious Posts Naver will disable comments on articles when malicious posts exceed a set threshold, as it steps up efforts to promote healthier online discussion. Naver said on 23 it is applying its artificial intelligence-based detection system, Cleanbot, to all articles, including those in the politics and elections sections, to identify malicious comments. When the number of malicious comments crosses the threshold, users will see a notice saying, “Comments are not available because Cleanbot detected a large number of malicious comments,” along with a banner for the company’s “Green Internet” campaign. First introduced in 2019 as an industry first, Cleanbot has been continuously upgraded to better detect profanity and sexually explicit or violent language, as well as expressions of hate, disparagement and discrimination, Naver said. The company added that it is preparing another AI Cleanbot model upgrade by the end of April. Kim Suhyang, a Naver leader, said the company has been working to make the comment section “a place for healthy communication” by advancing Cleanbot, following last month’s move to stop providing comments at the bottom of articles in the politics and elections sections. “We will continue to develop the service by listening to a range of opinions so we can respond to rapidly changing forms of malicious comments,” Kim said. * This article has been translated by AI. 2026-04-23 08:46:40
  • OpenAI Unveils ChatGPT Image 2.0 With Sharper Text and More Precise Layouts
    OpenAI Unveils ChatGPT Image 2.0 With Sharper Text and More Precise Layouts OpenAI on 22 announced a next-generation image model it said significantly improves output quality, widening the practical use of AI-generated images for workplace tasks. The company introduced the new image-generation model, called ‘ChatGPT Image 2.0.’ OpenAI said the model more precisely follows detailed user instructions, arranging objects and their relationships more naturally while improving accuracy in areas that have challenged earlier systems, including small text, icons, user interface elements and complex layouts. A key upgrade is text rendering. OpenAI said the model can produce readable lettering suited to real production needs such as posters, infographics and UI mockups, moving beyond simple image creation toward work-ready results. The model supports multiple aspect ratios, from up to 3:1 to 1:3, and can more precisely reproduce styles including photos, illustrations, comics and film looks. OpenAI said this enables outputs that can be used directly for proposals, marketing materials, educational resources and social content. OpenAI also said it strengthened multilingual performance, improving text rendering in languages including Korean, Japanese, Chinese, Hindi and Bengali for global content production. Users can generate up to 10 images at once to compare and review different options more quickly. OpenAI said the model is its first image generator to combine a ‘ChatGPT-based thinking’ function, allowing users to supplement information through web search or generate multiple images from a single prompt and check the results in a broader production workflow. ‘ChatGPT Image 2.0’ is available in ChatGPT and Codex. Advanced image-generation features are offered to subscribers on ChatGPT Plus Pro and business plans.* This article has been translated by AI. 2026-04-22 09:31:53
  • AhnLab Warns of Fake Claude Phishing Sites Spreading Info-Stealing Malware
    AhnLab Warns of Fake Claude Phishing Sites Spreading Info-Stealing Malware AhnLab said it has identified cases in which an information-stealing malware strain was distributed through phishing sites impersonating the generative AI service Claude, and urged users to be cautious. On 22, AhnLab said the phishing site displays the message “Bring Claude to your Desktop” and offers download buttons by operating system, including Windows and macOS, to lure users into installing it. When a user clicks the download button for their OS, a pop-up appears instead of an installer, providing installation instructions. The notice says the download will begin if the user copies a specific command and pastes it into their PC, but following those steps installs malware. Once installed, the malware steals files on the PC, browser-stored information and cryptocurrency wallet data, then sends it to an attacker-controlled server. AhnLab said this tactic — disguising instructions or error messages to get users to copy and paste a malicious command themselves — is known as the “ClickFix” technique and has recently been used in a range of malware distribution attacks. The company said the phishing site was found appearing near the top of Google search results for keywords such as “Claude app” and “Claude desktop.” It said attackers may have manipulated rankings by using search ads. AhnLab said such attacks continue to exploit users’ tendency to trust sites shown at the top of search results, and called for heightened vigilance. To reduce risk, it advised users to verify official domain addresses regardless of search rankings, apply the latest security updates for PCs, operating systems, software and browsers, and enable real-time monitoring in antivirus software such as V3. AhnLab said its V3 product line supports detection of the phishing site and blocks access, and that related threat information is available through its next-generation threat intelligence platform, AhnLab TIP.* This article has been translated by AI. 2026-04-22 09:15:07
  • NCSoft Confirms Global Launch of MMORPG Aion 2 in Second Half of 2026
    NCSoft Confirms Global Launch of MMORPG Aion 2 in Second Half of 2026 NCSoft has confirmed a global release in the second half of 2026 for its massively multiplayer online role-playing game, 'Aion 2.' The company said on the 22nd that it is developing the global version as a PC-only title for Steam and Purple. Players can add the game to their wishlists on Steam’s official page starting that day. NCSoft said it plans to build and operate servers by major regions, including North America, South America, Europe and Japan. The game will support 10 languages: English, German, French, Spanish, Portuguese, Japanese, Korean, Russian, and Chinese in both simplified and traditional forms. The Aion 2 development team plans to introduce the game in May and hold a session to address questions and communicate with global players. Baek Seung-wook, NCSoft’s chief business officer, said, "Aion 2 has continued to develop its content and services by listening to the community’s voice," adding, "We will provide these efforts and sincerity, community-centered operations, and a high level of game quality equally to users around the world." 2026-04-22 08:42:17
  • NHN Cloud, NC AI Partner on Cloud Infrastructure for Physical AI and Industry AI Shifts
    NHN Cloud, NC AI Partner on Cloud Infrastructure for Physical AI and Industry AI Shifts NHN Cloud and NC AI are joining forces to expand AI infrastructure-based business. NHN Cloud said April 22 that it signed a strategic memorandum of understanding with NC AI to jointly pursue business based on AI solutions and cloud infrastructure. The agreement is designed to combine the companies’ core technologies and business capabilities to support AI adoption across industries and strengthen their presence in South Korea’s cloud market. Under the partnership, the two companies will cooperate broadly on AI infrastructure operations and technology upgrades, joint sales and market expansion, and collaboration on government-funded projects and the public sector. They plan to provide AI infrastructure optimized for public and financial-sector customers, where security and stability are essential, and to bolster competitiveness through technical synergies. The companies also plan to work together so NC AI’s “world model” technology — which simulates real-world physical laws — and its “physical AI” technology, which interacts with real environments, can be deployed across industrial sites. NHN Cloud said it will provide a stable high-performance computing environment by optimizing computing resources based on its experience operating large-scale GPU infrastructure. “Ultimately, the foundation of AI competitiveness lies in stable, advanced infrastructure,” NHN Cloud CEO Kim Dong-hoon said. He said the partnership will support the deployment of physical AI and other AI services in industrial settings, build an AI transition environment that companies can adopt with confidence through industry-tailored services, and expand successful “AX” cases. NC AI CEO Lee Yeon-su said the core of the partnership is building an optimal infrastructure environment so the company’s physical AI and world model technologies can be applied immediately at industrial sites. He said the companies will use their combined strengths to drive successful “AX” for public and private-sector organizations. * This article has been translated by AI. 2026-04-22 08:23:23
  • Naver, Krafton and Mirae Asset Seek Up to 1 Trillion Won for India Tech Fund
    Naver, Krafton and Mirae Asset Seek Up to 1 Trillion Won for India Tech Fund Naver, Krafton and Mirae Asset are moving to make a large-scale technology investment push in India, aiming to identify promising information technology companies through a “Unicorn Growth Fund (UGF)” that could reach up to 1 trillion won. Naver said it introduced the UGF to local companies and venture capital firms at a briefing held April 21 in New Delhi. The event was arranged in connection with President Lee Jae-myung’s trip to India. Attendees included Industry and Trade Minister Kim Jung-kwan; Naver CEO Choi Soo-yeon; Krafton CEO Kim Chang-han; Krafton India head Son Hyun-il; Mirae Asset India Vice Chairman Swarup Mohanty; and Punit Kumar, CEO of Mirae Asset Venture Investment India. “India is rapidly expanding its AI industry on the back of abundant IT talent and a dynamic startup ecosystem, growing into a center of global digital innovation,” Choi said. She said the UGF will invest in high-growth technology companies in areas including AI, fintech and content, and serve as a platform to create strategic synergies by combining the three firms’ capabilities. Kim said India is emerging as the world’s fastest-growing gaming market and a hub for technological innovation. He said Krafton will be a mid- to long-term partner that supports promising companies’ global expansion through the UGF. Choi met April 20 with Commerce and Industry Minister Piyush Goyal to explain the fund’s vision and plans, and the two sides were reported to have formed a positive consensus. The UGF was formed early this year after Krafton committed 200 billion won, and began operations at more than 500 billion won including contributions from Naver, Mirae Asset and outside investors. The fund could later expand to as much as 1 trillion won, the companies said. The fund follows the Asia Growth Fund (AGF), jointly created by Naver and Mirae Asset. The AGF invested in major unicorns including India’s food delivery platform Zomato and Southeast Asia’s largest mobility company, Grab. Separately, Naver said it signed a memorandum of understanding the same day with Tata Consultancy Services, the IT arm of India’s Tata Group, to expand cooperation in AI, cloud and B2C services as it steps up its India strategy.* This article has been translated by AI. 2026-04-21 14:54:27