During the second day of the '2026 Taejae Future Education Forum,' held on October 21-22 at the International Conference Hall of the Seoul Chamber of Commerce and Industry, five education and edtech experts discussed the pitfalls of indiscriminately introducing AI technologies in universities and focused on innovative, learner-centered education and competency verification systems.
Won Do-yeon, Director of the AI Education Research Center at Taejae University: "Overreliance on AI leads to cognitive outsourcing; learners must be the agents of thought."
Won Do-yeon, Director of the AI Education Research Center at Taejae University, delivered the first presentation on 'Active Learning Expanded by AI: Keeping Learners as Agents of Thought and Judgment.'
He diagnosed that the paradox of automation seen during the Industrial Revolution and the ICT revolution of the 1980s is now being replicated in education as the 'paradox of AI augmentation.' While AI dramatically enhances cognitive productivity, it bypasses the unique pathways of human thought and reasoning, depriving learners of opportunities to accumulate their capabilities.
Won identified the main risks of AI overreliance in education as: ▲ cognitive outsourcing, where learners skip critical thinking; ▲ metacognitive illusions and authority bias, where learners blindly trust AI-generated answers; and ▲ cognitive constriction, where the breadth of thought narrows.
As an empirical alternative, he proposed the 'Inside Taejae Pre-College Education Model' principles. He stated, "In course design, AI should be limited to being a 'thinking partner.' When faced with a problem, the first question and initial judgment must come from the learner, and mechanisms for critical reflection on AI outputs should be integrated throughout the course."
In actual lesson plan examples, AI is restricted to specialized cognitive support roles, such as 'metacognitive monitoring mirror,' 'alternative scenario generator filling narrative gaps,' and 'visualization partner aiding creative embodiment.' Won emphasized the importance of course design that ensures learners' active cognitive efforts, noting that analysis of lesson data networks showed a strong correlation between students' reflective learning activities with AI and their achievement levels.
He diagnosed that the paradox of automation seen during the Industrial Revolution and the ICT revolution of the 1980s is now being replicated in education as the 'paradox of AI augmentation.' While AI dramatically enhances cognitive productivity, it bypasses the unique pathways of human thought and reasoning, depriving learners of opportunities to accumulate their capabilities.
Won identified the main risks of AI overreliance in education as: ▲ cognitive outsourcing, where learners skip critical thinking; ▲ metacognitive illusions and authority bias, where learners blindly trust AI-generated answers; and ▲ cognitive constriction, where the breadth of thought narrows.
As an empirical alternative, he proposed the 'Inside Taejae Pre-College Education Model' principles. He stated, "In course design, AI should be limited to being a 'thinking partner.' When faced with a problem, the first question and initial judgment must come from the learner, and mechanisms for critical reflection on AI outputs should be integrated throughout the course."
In actual lesson plan examples, AI is restricted to specialized cognitive support roles, such as 'metacognitive monitoring mirror,' 'alternative scenario generator filling narrative gaps,' and 'visualization partner aiding creative embodiment.' Won emphasized the importance of course design that ensures learners' active cognitive efforts, noting that analysis of lesson data networks showed a strong correlation between students' reflective learning activities with AI and their achievement levels.
Kim Sung-je, Chairman of Linus: "Mass personalization in public education; AI-LMS is the solution."
Kim Sung-je, Chairman of Linus and former Vice President of Hanyang Cyber University, presented on 'The Future of Education from Within Universities: Blueprint and Evidence of the AI-Connected University.' He sharply criticized the fundamental limitations of the modern public education structure, which originated from the 19th-century Prussian industrial model of 'mass schooling.'
He stated, "Despite students having vastly different aptitudes and learning speeds, the imposition of the same textbooks, pace, and exams has entrenched educational disparities. The most serious issue is the prevalence of 'measurable proxy' indicators that focus solely on easily quantifiable metrics like written exams or standardized tests." He explained that, similar to high TOEIC scores not guaranteeing practical English proficiency, the results-oriented mindset hidden behind safe test scores undermines the true purpose of education.
Kim identified 'AI-LMS' (AI-based Learning Management Systems) as the key to breaking this structural contradiction. He explained that while it is physically impossible for one teacher to cater to the needs of hundreds or thousands of students, the combination of AI-LMS and agent infrastructure can provide personalized feedback to tens of thousands simultaneously, enabling 'mass personalization' in education.
He proposed an 'authenticity architecture for public education' to ensure the integrity of learning, asserting that 'evaluation authenticity' and 'evidence-based assessment' are essential to demonstrate students' actual learning efforts beyond simple memorization tests. He added, "Instead of attempting to overhaul the entire university administration at once, we should first accumulate and disseminate 'model success' cases of classes combining one professor, a tutor, and an AI twin."
He stated, "Despite students having vastly different aptitudes and learning speeds, the imposition of the same textbooks, pace, and exams has entrenched educational disparities. The most serious issue is the prevalence of 'measurable proxy' indicators that focus solely on easily quantifiable metrics like written exams or standardized tests." He explained that, similar to high TOEIC scores not guaranteeing practical English proficiency, the results-oriented mindset hidden behind safe test scores undermines the true purpose of education.
Kim identified 'AI-LMS' (AI-based Learning Management Systems) as the key to breaking this structural contradiction. He explained that while it is physically impossible for one teacher to cater to the needs of hundreds or thousands of students, the combination of AI-LMS and agent infrastructure can provide personalized feedback to tens of thousands simultaneously, enabling 'mass personalization' in education.
He proposed an 'authenticity architecture for public education' to ensure the integrity of learning, asserting that 'evaluation authenticity' and 'evidence-based assessment' are essential to demonstrate students' actual learning efforts beyond simple memorization tests. He added, "Instead of attempting to overhaul the entire university administration at once, we should first accumulate and disseminate 'model success' cases of classes combining one professor, a tutor, and an AI twin."
Noh Won-seok, CEO of LECOS: "Transforming fragmented university experiences into evidence-based value; completing a four-year 'Odyssey' with open badges."
Noh Won-seok, CEO of LECOS, emphasized the need for a paradigm shift to support future education in his presentation titled 'Proving the Value of All University Experiences: Open Badges and Competency Ecosystems.'
He opened by stating, "Modern companies do not have the luxury of re-educating new hires in offline training centers for extended periods. They demand talent that can demonstrate 'what they can actually do' in the field, rather than graduates who have merely completed standardized credits in their major courses."
Noh identified three core values that future education must uphold in the AI era: ▲ Humanity (critical thinking, creativity, ethics); ▲ Learner Agency (choice, challenge, pathfinding); and ▲ Trust (evidence-based transparency). He pointed out that while students gain invaluable learning experiences outside of formal classrooms during their four years at university—through projects, internships, global challenges, research, and leadership—the existing grading system reduces these rich experiences to a few narrow grades, creating a contradiction of 'rich experiences but poor proof.'
As a solution, Noh proposed 'open badges' based on international technical standards. He explained that open badges are not just simple digital images or certificates but 'verifiable evidence packs' containing metadata such as the issuer, issuing organization, fulfilled achievement criteria, outputs, and observation records. He outlined a five-step roadmap for recognizing learners' experiences as valuable: 'Experience → Evidence → Competency → Credential → Opportunity.'
Noh suggested that without drastically changing existing legacy systems, universities should link distributed data from LMS and student information systems into an 'Evidence Hub' and integrate them into the international education technology standards of CASE (Competency Framework), Open Badges 3.0 (verifiable credentials), and CLR 2.0 (Comprehensive Learning Record). He added, "When we weave a student's four years into a narrative of 'Odyssey' from critical inquiry in the first year to AI collaboration in the second year, global challenges and applied practice in the third year, and systems leadership in the fourth year, learning becomes a complete lifelong asset for the student."
Noh also emphasized four principles for badge design: ▲ 'Evidence first,' strictly separating activities and competencies; ▲ 'Align to competency,' matching with core university competency standards; ▲ 'Open and portable,' not being dependent on specific platforms; and ▲ 'Human judgment,' where final decisions and accountability rest with educators, stating, "AI can recommend and interpret competencies, but final proof must come from reliable data and human evidence."
He opened by stating, "Modern companies do not have the luxury of re-educating new hires in offline training centers for extended periods. They demand talent that can demonstrate 'what they can actually do' in the field, rather than graduates who have merely completed standardized credits in their major courses."
Noh identified three core values that future education must uphold in the AI era: ▲ Humanity (critical thinking, creativity, ethics); ▲ Learner Agency (choice, challenge, pathfinding); and ▲ Trust (evidence-based transparency). He pointed out that while students gain invaluable learning experiences outside of formal classrooms during their four years at university—through projects, internships, global challenges, research, and leadership—the existing grading system reduces these rich experiences to a few narrow grades, creating a contradiction of 'rich experiences but poor proof.'
As a solution, Noh proposed 'open badges' based on international technical standards. He explained that open badges are not just simple digital images or certificates but 'verifiable evidence packs' containing metadata such as the issuer, issuing organization, fulfilled achievement criteria, outputs, and observation records. He outlined a five-step roadmap for recognizing learners' experiences as valuable: 'Experience → Evidence → Competency → Credential → Opportunity.'
Noh suggested that without drastically changing existing legacy systems, universities should link distributed data from LMS and student information systems into an 'Evidence Hub' and integrate them into the international education technology standards of CASE (Competency Framework), Open Badges 3.0 (verifiable credentials), and CLR 2.0 (Comprehensive Learning Record). He added, "When we weave a student's four years into a narrative of 'Odyssey' from critical inquiry in the first year to AI collaboration in the second year, global challenges and applied practice in the third year, and systems leadership in the fourth year, learning becomes a complete lifelong asset for the student."
Noh also emphasized four principles for badge design: ▲ 'Evidence first,' strictly separating activities and competencies; ▲ 'Align to competency,' matching with core university competency standards; ▲ 'Open and portable,' not being dependent on specific platforms; and ▲ 'Human judgment,' where final decisions and accountability rest with educators, stating, "AI can recommend and interpret competencies, but final proof must come from reliable data and human evidence."
Jeong Hoon, CEO of Learning Spark: "Universities must reflect on their AI readiness in supporting AX initiatives."
Another speaker, Jeong Hoon, CEO of Learning Spark, addressed the trends in IT services supporting university AX (digital transformation) and called for thorough self-assessment among higher education institutions.
He noted, "Future jobs are being restructured from simple repetitive tasks replaced by AI to complex problem-solving 'messy jobs,' prompting universities to question whether they possess the 'AI readiness' to meet these societal demands and digital transformations."
Jeong highlighted the importance of issues such as privacy, algorithm ethics, access gaps, data governance, and academic integrity within university communities, citing UNESCO reports. He cautioned against merely adopting commercial LLM tools, emphasizing the need for technical standards and data dictionaries to ensure interoperability among distributed academic and administrative legacy systems. He recommended a governance approach centered on continuous improvement and experimentation through 'pilot' projects rather than one-off large-scale initiatives to enhance overall execution in universities.
He noted, "Future jobs are being restructured from simple repetitive tasks replaced by AI to complex problem-solving 'messy jobs,' prompting universities to question whether they possess the 'AI readiness' to meet these societal demands and digital transformations."
Jeong highlighted the importance of issues such as privacy, algorithm ethics, access gaps, data governance, and academic integrity within university communities, citing UNESCO reports. He cautioned against merely adopting commercial LLM tools, emphasizing the need for technical standards and data dictionaries to ensure interoperability among distributed academic and administrative legacy systems. He recommended a governance approach centered on continuous improvement and experimentation through 'pilot' projects rather than one-off large-scale initiatives to enhance overall execution in universities.
Jeong Chan-pil, CEO of PickLab: "The combination of learning science and AI supports cognitive process design in edtech."
Jeong Chan-pil, CEO of PickLab, introduced the effects of integrating flipped learning and cognitive science theories in his presentation titled 'PICL: When Learning Science Meets AI.' A former documentary producer, Jeong shared compelling empirical cases where students who had previously been disengaged made significant progress when flipped learning was applied in the classroom.The '2026 Taejae Future Education Forum' was organized to explore the essential direction and core values for domestic higher education amid the global surge of generative AI. Over two days, university presidents, scholars, and global edtech leaders gathered to share concrete methodologies for integrating technology as a tool for achieving educational goals while guarding against technology-dependent education.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.