[Edaily Reporter Han Gwang-beom] K-AI models are rapidly making inroads into real-life settings, from streamlining national heritage administration to the digital transformation of offline financial branches.
On the 8th, the Ministry of Science and ICT unveiled additional examples of domestic AI models being applied in industrial settings. The announcement highlighted cases where Motif Technology and NC AI have integrated their proprietary AI technologies into public services and the financial sector, respectively, along with achievements in expanding the domestic AI semiconductor ecosystem and developing AI agents for vehicles.
In the fields of national heritage administration and public services, Motif Technology’s proprietary AI model is being introduced. The National Heritage Promotion Agency plans to utilize this to enhance administrative efficiency and lay the groundwork for providing prompt public services. In particular, by integrating this AI model into the currently operating national heritage image generation service "HAI," the agency aims to support the public in experiencing our heritage more creatively. "HAI" has also been selected as an official presentation case for the "AI for Good Global Summit" to be held in Geneva, Switzerland, this coming July.
Shim Jeong-taek, Head of the AI Data Team at the National Heritage Promotion Agency, stated, “Based on Motif’s AI technology, we plan to provide services that allow citizens to directly generate and utilize national heritage images, while also discovering various AI-based content and new services to broaden the scope of national heritage appreciation.” He added, “By applying AI technology to administrative services to provide swift and convenient public services, we will contribute to enhancing public welfare.”
In the financial sector, NC AI’s AI model is being deployed at Shinhan Bank’s financial branches to drive digital transformation. This involves building a “digital twin” environment that replicates the physical space of an offline branch in the virtual world, allowing for the pre-simulation of changes such as counter layout and kiosk configuration to support the design of an optimal financial environment. The goal is to provide customized spaces and services to customers visiting branches and to enhance the competitiveness of the financial industry.
Song Mu-kyung, a researcher at NC AI, said, “The financial sector has long relied on ‘human experience,’ and while counter layouts and customer flow patterns determine the customer experience, finding the optimal solution has always been hindered by barriers of time and cost.” He added, “This project is a promising attempt to break down those barriers, and we want to use AI and digital twin technology to systematically prove, through data, solutions to problems that were difficult to solve in the past.” He further emphasized, “Ultimately, our goal is for K-AI technology to contribute to transforming the entire financial industry and to create an experience where customers feel, ‘This space was designed just for me,’ the moment they walk through the branch door.”
The event also featured examples of AI infrastructure innovation through the integration of hardware and software. LG AI Research is collaborating with FuriosaAI on a “full-stack partnership” to run its proprietary super-large AI model, “EXAONE,” on FuriosaAI’s second-generation high-performance NPU, “RNGD (Renegade).” The two companies, which have been collaborating since 2023, plan to fully expand their services into the global AI ecosystem based on the synergy between the domestically developed NPU with improved power efficiency and their proprietary AI model.
In the mobility sector, domestically developed automotive AI is hitting the roads. 42dot recently unveiled its in-vehicle voice AI agent, 'Gleo AI,' showcasing a natural communication environment within the vehicle. Currently, SK Telecom and 42dot are collaborating to develop an in-vehicle LLM (Large Language Model) and are enhancing the AI interaction experience optimized for the driving environment by supporting the construction of specialized AI agent voice datasets.
