KT announced on the 30th that it held the 'Seoul National University-KT-KAIST Industry-Academia Joint Research Final Achievement Workshop' at the KT Umyeon Research Center in Seoul's Seocho District, together with researchers from Seoul National University and KAIST.
A total of over 50 researchers from KT, Seoul National University, and KAIST who attended this workshop shared the final results of their joint industry-academia research conducted over the past year and discussed how to apply these technologies at KT.
KT focused its research with Seoul National University on △context understanding and reasoning for autonomous agents, △RAI (Responsible AI) evaluation criteria and reliability improvement, △acquisition of Korean-specific datasets and development of behavior prediction models, and △user feedback reinforcement learning frameworks.
In addition, KT conducted joint research with KAIST on prompt compression and optimization technologies. This laid the groundwork for reducing the input processing burden of large language models and improving the inference efficiency of AI models and agent services.
KT plans to advance technical validation and internalization so that the core research achievements secured through this joint collaboration translate into actual model, platform, and service competitiveness. It also intends to identify follow-up projects reflecting business and technology needs to expand R&D cooperation in areas such as agentic AI, RAI, physical AI, and token efficiency.
Seoul National University AI Researcher Lee Jae-uk (Director) stated, "This joint research is significant because it connected the university's research capabilities with real-world technological challenges in industry and jointly verified the potential for applying research outcomes." He added, "Moving forward, we will continue to develop an industry-academia collaboration model through cooperation with KT that leads academic achievements into AI technology innovation and strengthened industrial competitiveness."
KAIST Graduate School of AI Dean Kim Jae-cheol Kim Gi-eung (Prof.) remarked, "The competitiveness of AI technology depends not only on model performance but also on efficiency and applicability in real-world environments." He expressed hope that "the technologies secured through this joint research will serve as a foundation for improving the processing efficiency of large language models and expanding the use of AI in industrial settings."
KT AX Future Technology Institute Director Park Jae-hyeong (SVP) said, "Through our collaboration with Seoul National University and KAIST, we have jointly completed the core AI technologies that KT requires." He further stated, "Building on this achievement, we will expand research and development cooperation models that grow together with universities and ensure that research outcomes continuously evolve into KT's AI competitiveness."
in TECH
