KT is partnering with Seoul National University and KAIST to develop new AI technologies, including physical AI. On September 3, KT announced that they hosted a kickoff workshop at their research center to expand AICT technology development cooperation. KT will provide the AI model “Trust” and commercialization capabilities, while the universities contribute deep research expertise to jointly develop key AI technologies such as autonomous agents, responsible AI (RAI), physical AI, human feedback-based reinforcement learning, and reasoning optimization.
Specifically, Seoul National University will focus on advancing AI core technologies that are human-centered and trust-based, including a robust context reasoning human feedback reinforcement learning framework for autonomous agents, RAI evaluation criteria, and trust improvement, as well as user-interaction actionable AI. KAIST will concentrate on developing core technologies that enhance the efficiency and accuracy of large-scale language models through prompt compression and optimization. KT plans to actively use the developed technologies to strengthen their AI lineup and apply them to various industries like public services, legal, finance, education, and healthcare, thereby building an AI cooperative ecosystem.
At the workshop, over 100 participants attended both online and offline, including KT’s Innovation Division head, Seung-pil Oh, and AI Future Lab head, Soonmin Bae, along with professors Byung-Tak Zhang from Seoul National University AI Research Institute and Gi-eung Kim from KAIST’s Kim Jae-chul AI Graduate School. Seung-pil Oh expressed expectations for tangible cooperation between KT and the university researchers, stating that the academic-industrial partnership model would have a positive ripple effect on the domestic AI industry. Professor Zhang emphasized the invaluable opportunity for AI technology development and talent cultivation through collaboration with KT’s extensive data and infrastructure. Professor Kim noted that the joint research serves as an exemplary cooperation model linking academia and industry, expected to drive future AI innovation and enhance industrial competitiveness.