This discussion, moderated by Hao Liu centres on how learning and education must be reimagined in the AI era, drawing on both European and Chinese traditions of knowledge acquisition.
Jovan Kurbalija of DiploFoundation opened by proposing that humans have historically learned through two fundamental methods: apprenticeship (learning by observing and doing) and storytelling . He argued that AI tools such as ChatGPT and DeepSeek now threaten the latter, as students can generate essays with minimal effort, raising the question of whether the ability to construct and tell one's own story remains important . Hong Guan of BIT's School of Global Governance responded by outlining five meta-capabilities considered essential for the AI era: learning agility, execution capability, communication skills, leadership potential, and critical judgement , emphasising that AI should improve education rather than replace it .
Student voices added practical perspectives. A Stanford University student expressed concern that classmates were uploading entire textbooks into AI tools to pass exams without genuinely learning, and questioned who would generate new knowledge if everyone depended on AI . Another student highlighted the risk of accepting AI-generated answers uncritically, arguing that students must develop the habit of questioning before consulting AI, not only after .
Senior practitioners from aviation and international organisations stressed that AI is a powerful tool but cannot substitute human vision, face-to-face interaction, or critical judgement . Ana Paula emphasised that adaptability and critical thinking are increasingly rare and valuable skills that can be deliberately cultivated . Maricela Munoz reinforced that distinctly human qualities - curiosity, compassion, and ingenuity - remain irreplaceable .
Jovan Kurbalija closed by reflecting that AI acts as a mirror revealing who we are, and invoked Pope Francis's encyclical Laudate Deum to argue that humans have a right to remain imperfect and should not be "optimised" by technology, a message he suggested resonates across philosophical and religious traditions worldwide .
Overall Purpose
- The discussion aims to explore how learning and education must evolve in the AI era, drawing on both European and Chinese traditions. Panellists, experts, and students examine the tension between AI as a productivity tool and the enduring importance of fundamentally human skills such as storytelling, critical thinking, and adaptability.
- --
Major Discussion Points
- Two foundational modes of learning - apprenticeship and storytelling - are being disrupted by AI. Jovan Kurbalija argues that throughout history, humans have learned in two core ways: by observing masters (apprenticeship) and by constructing and sharing narratives (storytelling) . He contends that AI now threatens the second mode, as students can generate polished essays with minimal effort, undermining the cognitive and emotional process of building a narrative .
- Students are using AI to pass without genuinely learning, raising urgent questions about the purpose of education. A Stanford University student described classmates uploading entire textbooks to AI tools the night before exams, passing with high marks but retaining nothing . This prompted the broader question of who will generate new knowledge in the future if students rely entirely on AI synthesis rather than original thought .
- Five meta-capabilities are proposed as essential for the AI era. Hong Guan outlined a framework developed at BIT's School of Global Governance: learning agility, execution capability, communication skills, leadership potential, and critical judgement . She argued that AI should not replace education but should push institutions to raise their standards, shifting the university's role from providing answers to helping students ask better questions .
- Critical thinking and adaptability are becoming increasingly rare and therefore increasingly valuable. Ana Paula observed that as access to information becomes cheap and ubiquitous through AI, the ability to evaluate, filter, and make sound decisions in a sea of contradictory data is now at a premium . She emphasised that adaptability is a learnable skill, not an innate trait, and that continuous learning is the quality most likely to sustain professionals through an unpredictable future .
- Human qualities - curiosity, compassion, and imperfection - remain irreplaceable and must anchor education in the AI era. Maricela Munoz stressed that humanity's inherent curiosity and ingenuity place humans in a strong position relative to machines, and warned against over-dependence on AI during hiring and professional development . Jovan Kurbalija echoed this in his closing remarks, referencing Pope Leo XIV's encyclical Dilexit Nos to argue that humans have a right to remain imperfect and that this imperfection is the source of life's ultimate meaning and beauty .
- --
Overall Tone
- The discussion maintains a consistently warm, intellectually curious, and collaborative tone throughout. The moderator, Hao Liu, sets an inclusive and informal atmosphere from the outset, encouraging open participation and limiting interventions to three minutes to keep energy high . Jovan Kurbalija's framing is reflective and philosophical, inviting challenge rather than asserting certainty . When students speak, the tone becomes more candid - but panellists respond with encouragement rather than alarm. Towards the close, the tone shifts gently towards the inspirational, with Maricela Munoz and Jovan Kurbalija offering humanistic reassurances about the enduring value of human qualities . Boris Engelson introduces a brief note of gentle scepticism and paradox , but this enriches rather than disrupts the overall spirit of open, good-natured inquiry.
Rethinking Learning for the AI Era: An Expanded Summary
#
Setting and Context
This discussion took place at what appears to be a WSIS (World Summit on the Information Society) international forum, moderated by Hao Liu. From the outset, Liu established an inclusive and informal atmosphere, inviting all participants to share ideas and reminding the room that every intervention should last no more than three minutes . He framed the session as a follow-up to a previous year's discussion on AI education comparing European and Chinese experiences , and introduced Jovan Kurbalija of the Diplo Foundation to explain the conceptual design of the session .
---
#
Two Foundational Modes of Learning
Jovan Kurbalija opened by proposing a deceptively simple but historically grounded framework: throughout human history, learning has relied on two core methods . The first is apprenticeship - learning by observing masters - which he argued underpins everything from the construction of the Great Wall of China to the Egyptian pyramids . The Diplo Foundation has translated this into what it calls "AI apprenticeship," where participants learn about AI by actively developing AI tools . The second mode is storytelling: the construction and sharing of narratives, whether around a fire, in a master's thesis, or in an academic article . Kurbalija acknowledged that his framing was rooted in Euro-Mediterranean tradition and explicitly invited Chinese participants to confirm or challenge whether these same modes apply in their context .
Kurbalija's central concern was that AI now threatens the second mode. He noted that platforms such as ChatGPT and DeepSeek can generate polished essays with minimal effort, meaning that a student who has been at a party until midnight can return home, enter a few prompts, and submit an excellent essay by morning . Rather than condemning this behaviour, he acknowledged it as natural: "if pedagogy is not associated with a common sense behaviour of people, it's not good" . His question, posed openly to the room, was whether the erosion of storytelling as a practised human skill constitutes a genuine problem . He argued that it likely does, because constructing a narrative requires the author to think about openings, tensions, and structure - cognitive processes that are bypassed when AI generates the text .
---
#
Storytelling as Motivation and the Chinese Tradition
Hao Liu responded by affirming that storytelling is critically important in Chinese history as well, confirming the cross-cultural resonance of Kurbalija's framework . He argued that throughout history, successful leaders have been skilled storytellers who built shared dreams, motivated followers without material reward, and made those dreams feel achievable . Crucially, he contended that this motivational power of storytelling cannot be extracted from AI tools: "you cannot take that from GPT, you cannot take that from DeepSeek" . This exchange produced one of the session's unexpected consensuses: despite Kurbalija's stated uncertainty about Chinese traditions , both Chinese and European participants converged on the view that apprenticeship and storytelling are universal human learning modes.
---
#
Five Meta-Capabilities for the AI Era
Hong Guan, representing BIT's School of Global Governance, introduced a structured framework of five meta-capabilities that her institution considers essential for the AI era, framing it as a "rebuilding" model designed to address the gap between existing educational approaches and the demands of an AI-driven world . The first capability is learning agility - the ability to learn how to learn, adapt quickly, and acquire new knowledge and skills in a rapidly changing environment . The second is execution capability - the ability to set goals, assess resources, make decisions, and deliver results, developed through real-world, project-based learning . The third is communication skills, understood not merely as writing and speaking correctly but as the ability to work effectively across different cultural backgrounds and with people from around the world . The fourth is leadership potential, defined not as occupying a high position but as taking responsibility, supporting teams, managing organisational relationships, and helping others move forward . The fifth and arguably most important is critical judgement - the capacity to evaluate information and make sound decisions in an AI-saturated world .
Hong Guan was clear about the underlying philosophy: "AI shouldn't replace education. AI should push us to make education better" . She argued that the role of universities must shift from providing answers to helping students ask better questions, exercise sound judgement, and act with responsibility .
---
#
Student Perspectives: Execution, Dependency, and Critical Reflection
The session then opened to student contributions, which added candid, experiential depth to the discussion. The first student, from BIT's School of Global Governance, focused on execution capability, arguing that in the AI era, idea generation has been largely commoditised: "everyone can use AI to generate a lot of solutions, ideas... within five minutes" . The real challenge, she argued, is turning AI-generated content into real-world practice . She described how her school's international study programmes - including visits to Geneva, Vienna, Egypt, and organisations such as the ITU and UNIDO - required students to develop complete execution plans with clear goals, resource assessments, and defined deliverables . This structured, project-based approach, she argued, produces genuine growth precisely because it is demanding .
The second student, a first-year undergraduate at Stanford University, offered the most emotionally resonant and intellectually challenging contribution of the session. She stated plainly that "AI is ruining education" , describing how classmates would upload entire textbooks into Claude or ChatGPT the night before exams, pass "with flying colors," and then "forget everything" the day after . She questioned the fundamental purpose of higher education under these conditions: "what's the point of being in school if you're just going to do this?" . She went further, making a pointed epistemological argument: ChatGPT does not understand what it says; it is "just a large language model that synthesizes everything that you put into it" and "doesn't understand a single thing" . Her deepest concern was not personal employment but systemic: "if every person at Stanford and every person at every other university is using AI, then who's going to generate this new knowledge for the future of humanity?" . She also noted the cultural dimension of the problem at elite institutions, observing that many of her peers were preoccupied with building AI-related startups rather than engaging in genuine intellectual inquiry . Her contribution was so striking that Kurbalija remarked warmly - in keeping with the session's informal atmosphere - that the Diplo Foundation might consider recruiting her, noting that they could not match a Stanford salary .
A third student raised a subtler but equally important concern about critical judgement. They observed that when students ask AI a question they do not fully understand, the AI's response often gives them a false sense of comprehension, causing them to stop asking further questions . Because they lack the background knowledge to evaluate the answer, they absorb it passively without genuine understanding . Their proposed remedy was practical and metacognitive: students should engage in critical reflection before consulting AI, clarifying precisely what they are confused about so that they can evaluate the AI's response with greater discernment . A fourth student extended the communication theme, arguing that AI has not weakened the importance of communication competence but has in fact strengthened it, since effective human-AI interaction now requires the ability to formulate precise, well-structured instructions .
---
#
Institutional Responses: Oral Examinations and Assessment Reform
The Stanford student's account prompted an immediate institutional response from Hong Guan, who contrasted BIT's approach with the written assessment model that appears to prevail at Stanford. She stated that BIT does not use the traditional examination model; instead, it employs oral examinations conducted across multiple rounds of questioning . This format, she argued, makes AI-assisted cheating effectively impossible, since genuine understanding is revealed - or exposed as absent - in real time . Jovan Kurbalija validated this approach by asking the Stanford student directly whether she had actually read the 500-page textbook , and she confirmed that she had - implicitly illustrating that direct questioning distinguishes genuine learners from those who rely on AI. Hao Liu reinforced the principle during the session itself by asking students to speak without manuscripts, noting that this reveals whether the ideas are genuinely their own .
Kurbalija expressed broader concern about the cultural environment at elite universities, describing it as "very worrying if at Stanford University you have this startup obsession, not people reflecting" . He argued that university should be "a place where you spend free time... to think, to reflect, to stay back, to have like-minded people, to engage, to question your thoughts" , and that the persistence of old assessment models was driving the problem . Hao Liu also floated an informal, off-the-cuff proposal for a two-tier credentialing system: one degree category indicating AI-assisted completion and another indicating independent completion, allowing AI use to be acknowledged transparently rather than prohibited or ignored .
---
#
Perspectives from Aviation and International Organisations
Two senior aviation experts offered perspectives grounded in professional practice. Catalin Radu, introduced by Hao Liu as a former Director General of Civil Aviation for two countries and a former ICAO official, framed AI unambiguously as "a tool to be used by human in a proper way to get better results, more effective, more productive" . He noted that AI has already transformed his own working practice, reducing the need for multiple administrative assistants . However, he emphasised that human vision, critical thinking, and face-to-face interaction remain irreplaceable, particularly for sensitive strategic discussions . His most urgent message was directed at organisations: AI adoption must happen system-wide and immediately. "It should happen now, today. So it's already late," he stated, arguing that treating AI as the responsibility of a single designated individual is wholly insufficient . He described AI readiness not as optional but as "mandatory" for organisational survival .
Nabil Naoumi, introduced by Hao Liu as the youngest president of ICAO's Air Navigation Commission - described as the highest technical body in that organisation - offered a complementary perspective. He argued that AI's greatest practical value lies in its ability to produce drafts and compile information rapidly, allowing expert groups to focus their energy on refinement and decision-making rather than information gathering . He echoed Radu's urgency: "if you just start now to think about AI, you are already too late. It is part of our daily business" . He drew a historical parallel with the arrival of the internet 25 years ago , suggesting that adaptation to transformative technology has always been necessary. He also illustrated how deeply AI has already penetrated aviation: aircraft systems are now capable of autonomous emergency landings, communicating with air traffic control, and safely landing a plane if a pilot becomes incapacitated . He noted that decision-makers in aviation are actively considering replacing the two pilots in a cockpit with AI systems, given the growing complexity of aircraft and the global shortage of pilots .
---
#
Critical Thinking, Adaptability, and Lifelong Learning
Ana Paula, who founded AIM Global at UNIDO and also served as Chief of Digital Innovation and AI before taking up her current role as Chief of ICT, brought a practitioner's perspective shaped by decades of technological change. She began by noting that the specific knowledge she studied - electronics engineering, programming in assembly language - is largely obsolete today , and that what has sustained her career is not any particular technical expertise but a commitment to lifelong learning established in her early teens . Her core argument was that as AI makes information cheap and ubiquitous, the ability to evaluate, filter, and make sound decisions in a sea of contradictory data is becoming increasingly rare and therefore increasingly valuable: "critical thinking is coming at a premium now because information is now cheap" . She also challenged the assumption that adaptability is an innate trait, arguing that it is a learnable skill that can be deliberately cultivated and improved over time . She noted that learning itself is evolving in the AI era: when teachers expect students to use AI, the bar is raised considerably higher, requiring deeper engagement rather than less . Her closing message was that "continuous learning" and adaptability are "the skills that will propel into the future, whatever that is" .
---
#
Human Qualities and the Irreplaceability of Authentic Humanity
Maricela Munoz, a diplomat with experience in international science and technology governance and a participant in the Diplo Foundation's apprenticeship programme - alongside other colleagues such as Leonardo - offered a humanistic perspective that complemented and deepened the preceding contributions. She argued that humanity's inherent curiosity and ingenuity place humans in a fundamentally strong position relative to machines: "our humanity and our curiosity will be always above any machine who can imitate, for example, critical thinking" . She noted that in professional contexts, she actively looks for candidates who remain "very human and very natural" and use technology as a tool rather than depending on it . Her framing echoed Radu's earlier point that AI is ultimately a tool in service of human agency, though she approached it from a more humanistic angle. She cited the singer John Legend's observation that technology cannot know "what love is, what pain is, and so many other human characteristics" , arguing that the creation of knowledge, wisdom, and unique vision "remains a quality and a project and a vision for a human being" . Her practical recommendation was that technology should be used to free humans from mundane tasks, creating more time for reflection, vision, and creative thinking . She closed with an encouragement to the younger participants to "continue being us, fresh," affirming that their critical thinking and human qualities are already present and valuable .
---
#
The Contrarian Voice: Persistence of Tradition and Unresolved Paradoxes
Boris Engelson, described by Kurbalija as the Diplo Foundation's "resident contrarian," introduced a layer of philosophical scepticism that enriched the discussion considerably. He observed that dissatisfaction with schooling is not a modern phenomenon: "one of the first texts of the history of writing, dating back to Sumer, was a complaint of a pupil" . Despite centuries of reform calls, "the format remains nearly intact through the century and even with the most modern technologies" . He noted with gentle irony that the very session discussing educational reform was itself conducted in a traditional lecture-panel format , suggesting that there must be "deep reasons why this model cannot change" . He also articulated a second, unresolved paradox: the tension between the Renaissance ideal of total encyclopaedic knowledge - embodied by Pico della Mirandola, whom Engelson described as a Renaissance man who "knew everything" - and the Einsteinian ideal of radical disruption . Both models are necessary, he argued, yet they are fundamentally difficult to reconcile within a single educational system. He concluded by noting that he prefers to live with unanswered questions rather than pretend to have answers , a stance that added intellectual honesty to the discussion.
Kurbalija, in introducing Engelson, also reflected on the limits of common sense as a guide for critical thinking. He noted that Engelson had taught him that "common sense has one dangerous element - that is a common, and very often common sense is not critical thinking, it's just very often cacophony" . This created a productive tension with Kurbalija's own earlier argument that pedagogy must align with common-sense behaviour , suggesting that the relationship between pragmatic realism and genuine critical inquiry is itself complex and unresolved.
---
#
Closing Reflections: Technology as Mirror and the Right to Imperfection
Jovan Kurbalija's closing remarks synthesised the session's themes into a broader philosophical reflection. He introduced the concept of "Esprit Tech de Genève" - a fan featuring eight philosophers, including Borges, Calvin, Rousseau, Voltaire, Piaget, Bonnet, and others - as a symbol of the idea that every place has its own philosophical tradition relevant to navigating the AI era . He encouraged participants to develop their own localised "Esprit Tech," drawing on the wisdom of Confucius, Lao Tse, African Ubuntu, and other traditions, rather than defaulting to a Western-centric framework .
His most resonant observation was that AI "puts a mirror in front of us to see who we are" - a metaphor suggesting that the discomfort AI causes is not merely a technical or pedagogical problem but an invitation to deeper self-understanding. He invoked a papal encyclical - which he described informally as a "Magna Carta of our era" and referred to as "Humanitas" - to make the case that humans have "a right to be imperfect" and that "in our imperfection is our ultimate purpose and beauty of the life" . He argued that this message resonates across philosophical and religious traditions from Asia to Africa and Latin America, because it is "centered in the role of humans, what it matters to be human" .
---
#
Practical Initiatives and Forward-Looking Commitments
Before closing, Hao Liu drew attention to the AI Philosophy Caravan initiative, which offers participants an immersive, multi-city learning experience engaging with government officials, industry representatives, academics, and local communities to understand real-world AI governance models across sectors such as agriculture and manufacturing . He noted that the upcoming 2026 edition would adopt a modified format - establishing a base in one city and inviting people to travel to meet the group, rather than requiring all participants to travel continuously - in response to feedback from previous participants . Participants were encouraged to follow the organising team on LinkedIn for updates . Hao Liu also took a moment to thank Sorina, described as the person "standing behind the curtain, making so many things happen," in recognition of the behind-the-scenes work that made the session possible.
---
#
Overall Assessment
The discussion achieved a notably high degree of cross-cultural and cross-generational consensus on several core themes: that AI is a tool which must remain in service of human agency; that critical thinking and adaptability are the most valuable skills in an AI-saturated world; that distinctively human qualities - storytelling, curiosity, compassion, and the capacity for genuine reflection - cannot be replicated by machines; and that Chinese and European learning traditions share the same fundamental modes of apprenticeship and storytelling. The most productive tensions emerged between the student voices - particularly the Stanford student's frank account of AI undermining genuine learning - and the professional voices of Radu and Naoumi, who enthusiastically endorsed rapid AI adoption . Boris Engelson's contrarian observations about the structural persistence of traditional schooling and the paradox between encyclopaedic knowledge and disruptive innovation added philosophical depth that prevented the discussion from settling into easy optimism. The session moved from diagnosis through institutional response to humanistic synthesis, leaving participants with both practical frameworks and enduring questions about the nature and purpose of learning in the age of AI.
Two core learning methods: observing/apprenticeship and storytelling have existed since time immemorial - Learning by observing and narrating
Arg. 1Jovan Kurbalija argues that humanity has always learned through two fundamental methods: apprenticeship (learning by observing a master) and storytelling (conveying knowledge through narratives). He contends that everything from the Great Wall of China to the pyramids was built through apprenticeship, while storytelling encompasses everything from fireside tales to academic dissertations. AI now threatens the storytelling method by making it easy to generate essays and narratives without genuine engagement.
Kurbalija illustrated apprenticeship by referencing how great historical achievements such as the Great Wall of China and the pyramids were built by people observing their masters , and noted that Diplo uses 'AI apprenticeship' where participants learn about AI by developing AI . He described storytelling as ranging from telling stories around a fire to writing a master's thesis or academic article , and observed that AI tools like DeepSeek and ChatGPT now allow students to generate essays overnight without genuine effort .
on: Human qualities — curiosity, compassion, ingenuity, and the capacity for storytelling — remain irreplaceable by AI
on: Whether the loss of storytelling and essay-writing skills is a serious problem
Pedagogy must reflect reality - Pedagogy must align with common-sense human behaviour; if students have access to AI tools, they will naturally use them, making it futile to ignore this reality
Arg. 2Kurbalija argues that educational systems must acknowledge the reality that students will use whatever tools are available to them, including AI, rather than pretending those tools do not exist. He contends that if pedagogy is disconnected from common-sense human behaviour, it is ineffective. This means educators must redesign learning rather than simply prohibiting AI use.
He gave the example of a student who has been at a party until midnight and must submit an assignment the next day - of course they will use DeepSeek or ChatGPT to write the essay . He stated plainly that 'if pedagogy is not associated with a common sense behavior of people, it's not good' .
on: AI undermines genuine learning when used to bypass the process of knowledge acquisition
on: Whether common sense supports or undermines critical thinking
University culture and reflection - The startup obsession at elite universities risks crowding out genuine reflection and the deeper purpose of higher education as a space for thinking and questioning
Arg. 3Kurbalija expresses concern that universities, exemplified by Stanford, are becoming dominated by a startup culture focused on AI chatbots and commercial ventures rather than fostering deep reflection and intellectual questioning. He argues that the true purpose of university is to provide free time and a community of like-minded people for thinking and questioning one's own thoughts. This cultural shift represents a worrying departure from the foundational mission of higher education.
Responding to Student 2's account of Stanford, Kurbalija stated that 'university is a place where you spend free time... to think, to reflect, to stay back, to have like-minded people, to engage, to question your thoughts' . He noted it was 'very worrying if at Stanford University you have this startup obsession, not people reflecting' , and acknowledged this problem is 'not obviously... an example from Stanford, but it is all over the place' .
on: Whether traditional educational assessment models are adequate in the AI era
Technology as a mirror for humanity - Technology acts as a mirror, prompting humanity to reflect on who we are, and we must accept our imperfections rather than seek technological optimisation
Arg. 4Kurbalija argues that one of AI's most valuable contributions is that it holds up a mirror to humanity, forcing us to confront and reflect on who we truly are. He draws on Pope Francis's encyclical Laudate Deum (Humanitas) to argue that humans have a right to be imperfect and should not be optimised by technology. Our imperfection, he suggests, is the source of our ultimate purpose and the beauty of life.
Kurbalija referenced Pope Francis's encyclical, paraphrasing its message as urging people to 'come down, we were thinking for 20 centuries and don't be too excited about technology' , and highlighted the Pope's argument that 'we have a right to be imperfect and in our imperfection is our ultimate purpose and beauty of the life' . He noted this message resonates across all thinking and religious traditions from Asia to Africa and Latin America .
on: Human qualities — curiosity, compassion, ingenuity, and the capacity for storytelling — remain irreplaceable by AI
Global philosophical traditions - The philosophical traditions of diverse civilisations, from Confucius to African Ubuntu, offer enduring wisdom relevant to navigating the AI era
Arg. 5Kurbalija argues that every civilisation has developed its own philosophical tradition for reflecting on the human condition, and these traditions remain deeply relevant in the AI era. He points to the concept of 'Esprit Tech' — the intellectual spirit of a place — as something that can be drawn from any culture, whether Geneva's Enlightenment thinkers, Confucius and Lao Tse in China, or Ubuntu philosophy in Africa. These traditions provide enduring frameworks for understanding purpose, dignity, and meaningful life.
Kurbalija described the 'Esprit Tech de Genève' concept, referencing philosophers such as Borges, Calvin, Rousseau, Voltaire, Piaget, and Bonnet , and noted that China has its own equivalent in Confucius and Lao Tse, referencing a philosophy tour in China where colleagues discussed their legacy . He also cited the African Ubuntu civilisation as an example of a tradition where people were 'thinking and reflecting on society' .
Storytelling as natural motivation - Storytelling is a powerful motivational tool that drives people toward shared dreams without material rewards
Arg. 1Hao Liu argues that storytelling is a critically important capability throughout human history, as all successful leaders have used it to build shared visions and motivate people to work toward common goals. He contends that storytelling achieves this motivational effect without any material incentives — no medals, gold, or silk — making it a uniquely powerful human tool. This capability, he argues, cannot be replicated by AI tools such as GPT or DeepSeek.
Liu observed that 'all of the successful people are good storytelling people' throughout history , explaining that they build a big dream and then share stories to make people believe the dream will come into reality and motivate them to work toward it . He emphasised that this motivation works 'without paying them anything. Not a medal, not the gold, not the silk' , and concluded that 'you cannot take that from GPT, you cannot take that from DeepSeek' .
on: Human qualities — curiosity, compassion, ingenuity, and the capacity for storytelling — remain irreplaceable by AI
on: Whether the loss of storytelling and essay-writing skills is a serious problem
Five meta-capabilities framework - Five meta-capabilities critical for the AI era: learning agility, execution capability, communication skills, leadership potential, and critical judgment
Arg. 1Hong Guan presents a framework of five meta-capabilities that her institution, BIT's School of Global Governance, has identified as essential for students in the AI era. These capabilities — learning agility, execution capability, communication skills, leadership potential, and critical judgment — are designed to address the gap between existing educational models and the demands of the AI era. The framework is grounded in the recognition that every country already has successful learning models, but these may not be adequate for the new context.
Hong Guan explained that learning agility involves knowing how to learn and adapt quickly to new situations , execution capability means setting goals, using resources, making decisions, and delivering results through real projects , communication skills encompass working across different backgrounds and cultures , leadership potential involves taking responsibility and supporting teams , and critical judgment means developing the ability to think critically about everything in the AI world .
on: Adaptability and continuous learning are essential capabilities for navigating an unpredictable technological future
AI as education enhancer - AI should not replace education but push universities to help students ask better questions and act with responsibility
Arg. 2Hong Guan argues that the role of universities in the AI era should shift from providing answers to helping students formulate better questions, exercise sound judgement, and act responsibly. She contends that AI should be seen as a catalyst for improving education rather than a substitute for it. This reframing positions the university as a facilitator of critical and responsible thinking rather than a repository of knowledge.
She stated clearly that 'AI shouldn't replace education. AI should push us to make education better' , and described the new role of universities as helping students 'ask a better question, make better judgment and act with the responsibility' .
on: AI is a tool to enhance human capability, not a replacement for human thinking and judgement
on: Whether AI is fundamentally ruining education or merely transforming it
Oral exams as AI-proof assessment - Oral examinations and project-based assessments can effectively distinguish genuine learners from those who rely on AI to cheat
Arg. 3Hong Guan argues that BIT avoids the problem of AI-assisted cheating by using oral examinations rather than written essays, making it impossible for students to rely on AI-generated content in the moment of assessment. She contends that oral exams, with multiple rounds of questioning, allow educators to evaluate genuine understanding and distinguish authentic learners from those who have used AI to cheat. This approach is presented as a practical institutional solution to the academic integrity challenge posed by AI.
She explained that BIT uses oral examinations where students are questioned in multiple rounds and evaluated on their real-time performance , stating 'we know how to torture you, we know how to challenge you' . She asserted that under this model, 'those are people who are really reading that 500 pages' and can be distinguished from those 'only using the large model language to help to cheat' .
on: AI undermines genuine learning when used to bypass the process of knowledge acquisition
on: Whether traditional educational assessment models are adequate in the AI era
AI enabling superficial learning - Students using AI tools to pass exams without genuinely learning the material, raising questions about the purpose of education
Arg. 1Student 2, a first-year student at Stanford University, argues that AI is undermining genuine education by enabling students to pass exams without truly learning the material. She describes classmates who upload entire textbooks and course materials to AI tools the day before exams, pass with high marks, and then forget everything the next day. This raises a fundamental question about the purpose of attending university if students are not genuinely acquiring knowledge.
She described classmates in her linear algebra class (Math 51) who, instead of reading the 500-page textbook, would upload it to Claude or ChatGPT and cross-reference it the day before lectures . She noted that 'the day before exams, my classmates would upload the whole textbook and the entire course material and all the lecture videos onto Claude, and then they would pass the test with flying colors. But then the day after the exam, they forget everything' .
on: AI undermines genuine learning when used to bypass the process of knowledge acquisition
on: Whether traditional educational assessment models are adequate in the AI era
Threat to future knowledge generation - The concern that if everyone relies on AI, no one will generate new knowledge or make genuine academic discoveries
Arg. 2Student 2 raises a systemic concern that if all students at elite universities rely on AI to generate knowledge, there will be no one left to produce genuinely new research, discoveries, or ideas for the future of humanity. She points out that AI itself only knows what humans have put into it, meaning it cannot generate truly novel knowledge. The preoccupation with AI startups further compounds this concern by diverting talented students away from substantive intellectual contribution.
She argued that 'AI doesn't understand what it's saying' and is 'just a large language model that synthesizes everything that you put into it' , questioning 'if every person at Stanford and every person at every other university is using AI, then who's going to generate this new knowledge for the future of humanity?' . She also expressed concern about 'who's going to be writing all these new research papers, who's going to be making new discoveries' if everyone is focused on AI startups .
on: How urgently organisations must adopt AI
Execution over idea generation - AI can generate ideas and solutions rapidly, but the real challenge is turning AI-generated content into real-world practice
Arg. 1Student 1 argues that while AI has made the generation of ideas and solutions trivially easy, the genuine challenge — and therefore the most valuable skill — lies in translating those AI-generated outputs into real-world practice. In the past, the scarcity was in ideas and knowledge; now the scarcity is in execution. This makes execution capability the critical differentiator in the AI era.
She noted that 'in the past we may lack like knowledges, ideas, solutions, but now everyone can use AI to generate a lot of solutions, ideas, like those things within five minutes or less than five minutes' , and posed the key question: 'how can we turn the AI-generated contents into real practice?' .
Project-based learning for execution - Project-based learning with clear goals, resource assessment, and defined deliverables builds genuine execution capability in students
Arg. 2Student 1 describes how her school's approach to building execution capability involves practical, project-based learning through visits to international organisations, combined with rigorous pre-departure planning exercises. Students are repeatedly challenged by professors to define their goals, assess available resources, and specify the results they intend to deliver. This structured approach to real-world engagement builds genuine execution skills.
She described her school's winter and summer semesters, during which students travel to Geneva, Vienna, Egypt, and international organisations such as ITU and UNIDO for practical study . She explained that before each departure, professors constantly challenge students with questions about goals, resources, and desired results , and that having 'a complete and clear execution plan' is required, which she described as 'really tough' but productive of 'real growth' .
Expanded communication competence - Communication competence now extends beyond human-to-human interaction to include precise instruction-giving to AI systems
Arg. 1Student 4 argues that the emergence of AI has expanded the scope of communication competence beyond traditional human-to-human interaction to include the ability to communicate effectively with AI systems. Giving precise instructions to AI — crafting prompts that yield useful outputs — is itself a form of communication that requires skill and clarity. Rather than weakening the importance of communication, AI has actually strengthened and broadened it.
She observed that before AI, communication meant speaking clearly and listening carefully to others, but now it also requires 'thinking about the words we key in the chat box and the precise instructions that we give' to AI so that it produces outputs that meet expectations . She concluded that 'AI didn't weaken the communication competence importance but also like strengthen it' .
AI as professional tool - AI is a tool to be used by humans to achieve better, more effective, and more productive results, not a replacement for human vision and judgement
Arg. 1Catalin Radu argues that AI is fundamentally a tool — a powerful one — but that it must be used by humans in a proper way to enhance results rather than replace human thinking and vision. He emphasises that while AI can handle many administrative and informational tasks, the strategic vision, critical thinking, and face-to-face human interaction required for complex decision-making remain irreplaceable human contributions. The key is learning how to use the tool effectively.
He stated that AI 'is indeed a tool to be used by human in a proper way to get better results, more effective, more productive' , and gave the personal example that he no longer needs to hire two or three assistants for bookkeeping and administrative tasks . However, he stressed that 'in critical thinking, in the way you see the vision for your business or for your future challenges... you need to think clear' , and noted that face-to-face interaction remains 'extremely important' for sensitive business matters .
on: AI is a tool to enhance human capability, not a replacement for human thinking and judgement
on: Whether the loss of storytelling and essay-writing skills is a serious problem
Urgent organisational AI adoption - Organisations must implement AI system-wide and change their processes now, as those who have not started are already behind
Arg. 2Radu argues that organisations cannot afford to delay AI adoption and must implement it comprehensively across all processes and departments, not merely designate a single 'AI person'. He contends that it is already too late for organisations that have not begun this transformation, and that AI readiness is now mandatory for survival in the coming era. The change must be systemic, affecting processes and models throughout the entire organisation.
He warned that 'a lot of people don't realize nowadays that it's the moment that you should implement AI in your company, in your business, in your organization, and not just to have one guy saying he's the AI guy in the organization' . He stated that organisations need to 'have the whole system change' and that 'it should happen now, today. So it's already late' , adding that 'it's mandatory, you know. You cannot survive the next era without starting to implement and to modify the processes inside your organization' .
on: Adaptability and continuous learning are essential capabilities for navigating an unpredictable technological future
on: How urgently organisations must adopt AI
AI for efficient drafting - AI saves significant time by producing drafts and compiling information rapidly, allowing expert groups to focus on refinement and decision-making
Arg. 1Nabil Naoumi argues that one of AI's greatest practical contributions in professional settings is its ability to rapidly produce draft documents and compile information, which previously required significant human effort and time. In expert group settings such as ICAO meetings, someone must always prepare a draft before substantive discussion can begin; AI can now produce this in seconds. This allows human experts to focus their energy on refinement, deliberation, and decision-making rather than information gathering.
He described the ICAO meeting process, where a group of experts from different nations must have a draft prepared before they can begin working, noting that 'someone has to do a draft, first of all, to collect all possible information and present it to you, and then you start working on it' . He observed that 'with the AI, you can have the draft within minutes, within seconds, and then you start working on it' .
on: AI is a tool to enhance human capability, not a replacement for human thinking and judgement
on: How urgently organisations must adopt AI
AI in aviation operations - In aviation, AI is already embedded in aircraft systems capable of autonomous emergency landings, illustrating how deeply AI has penetrated critical professional domains
Arg. 2Naoumi illustrates the depth of AI's integration into professional domains by pointing to aviation, where AI systems are already capable of autonomously managing emergencies, including landing aircraft without pilot input. This example demonstrates that AI is not merely a future prospect but an operational reality in safety-critical industries. It also raises profound questions about the future role of human professionals in domains where AI can perform core functions.
He described emergency systems already on the market that can land an aircraft autonomously if the pilot is incapacitated - the aircraft detects a potential medical emergency, takes over communications, changes direction and altitude, and lands safely . He also noted that modern aircraft can refuse pilot commands if the AI determines the instruction is unsafe, stating 'you will give an order to the airplane to climb or to descend but the airplane may refuse' . He mentioned that the aviation industry is considering replacing the two pilots in the cockpit with artificial intelligence .
Premium on critical thinking - Critical thinking is becoming increasingly precious and rare as information becomes cheap and ubiquitous through AI
Arg. 1Ana Paula argues that as AI makes information cheap, abundant, and easily accessible, the ability to think critically — to evaluate, synthesise, and make sound judgements from a sea of information — becomes increasingly scarce and therefore increasingly valuable. The traditional expert power derived from holding exclusive access to information has already been disrupted. Critical thinking is now the premium skill that differentiates effective professionals.
She stated that 'critical thinking is coming at a premium now because information is now cheap. It comes from all over. AI knows more than you could ever' , and observed that 'the notion of holding power, expert power through information, yes, that's changed already' . She emphasised that 'you need to know what is correct, you need to know how to make a decision in a sea of information and contradictory information, and everybody is proclaiming to be an expert' .
on: Critical thinking is the most valuable and increasingly scarce skill in the AI era
Adaptability as a learnable skill - Adaptability and continuous learning are learnable skills, not innate traits, and are essential for navigating an unpredictable future
Arg. 2Ana Paula argues that adaptability is not an innate personality trait but a skill that can be deliberately learned and improved over time. She draws on her own career trajectory — from electronics engineering to data science and AI policy — to illustrate that the specific knowledge one studies often becomes obsolete, making the capacity to adapt and continue learning the most durable professional asset. Continuous learning and resilience are what carry professionals forward through technological change.
She reflected on her own career, noting that what she studied first 'practically doesn't exist anymore' - she began as an electronics engineer working with discrete components and programming in assembly - and that 'what you think you're going to be working 10 years, you're probably wrong' . She stated that 'adaptability is actually a skill. It can be learned. It can be improved. It's not an innate, it's not something that you were born with' , and noted she is currently on her third degree and still learning .
on: Adaptability and continuous learning are essential capabilities for navigating an unpredictable technological future
on: Whether AI is fundamentally ruining education or merely transforming it
Lifelong learning commitment - A commitment to lifelong learning, established early in life, is the most durable foundation for a career, as the specific knowledge studied often becomes obsolete
Arg. 3Ana Paula argues that the most important thing a young person can cultivate is not a specific body of knowledge but a deep commitment to continuous learning, because the specific content of any degree is likely to become obsolete. She draws on her own experience of making this commitment at age 13 or 14, which she credits as the foundation that has sustained her through multiple career transitions and technological shifts. The commitment to learning itself, rather than any particular expertise, is what endures.
She described making a commitment at around age 13 or 14 - 'I love studying, I'm going to study and learn for the rest of my life' - and credited this as what 'kept me going because what I studied first practically doesn't exist anymore' . She noted she is currently pursuing her third degree and is 'still studying' , and affirmed that 'continuous learning' is among 'the skills that will propel into the future, whatever that is' .
on: Adaptability and continuous learning are essential capabilities for navigating an unpredictable technological future
False sense of knowledge from AI - AI generates answers that make students feel satisfied, causing them to stop asking further questions and preventing deep understanding
Arg. 1Student 3 argues that a subtle but significant danger of AI is that it produces answers that give students a false sense of understanding, causing them to stop their inquiry prematurely. Because students lack the background knowledge to evaluate whether an AI's answer is accurate or reliable, they accept it uncritically and fill their knowledge gaps with AI-generated content they do not truly understand. This prevents the development of genuine comprehension and personal judgement.
She described the phenomenon where students ask AI a question they are unsure about, and 'what it gives us may mostly make us feel, oh, I may know the answers. And so we stop asking the further questions' . She explained that 'when we ask about questions, we don't know the relative knowledge. So we have no capacity or enough capability to make sure whether this answer is reliable, accuracy' , resulting in students 'filling the knowledge, but cannot fully understand the meaning of it or have some personal judgment' .
on: AI undermines genuine learning when used to bypass the process of knowledge acquisition
Pre-AI critical reflection - Before using AI, students should first clarify their own understanding and questions so they can critically evaluate the AI's response
Arg. 2Student 3 proposes a practical strategy for maintaining critical thinking in the age of AI: students should engage in reflective questioning before turning to AI, clarifying exactly what they are confused about and what they already know. This pre-AI reflection equips students with enough context to evaluate the AI's response critically rather than accepting it passively. She argues that critical practice must be extended to the moment before AI use, not only applied after receiving a response.
She proposed that 'we have to make the critical practice be longer, which means not only practice after using the AI, we have to make it before using it' . She suggested that 'before using AI, we can ask ourselves questions, what is exactly we are confusing about? When we know it, we can have our maybe better judgment about the response and have the critical thoughts' .
on: Critical thinking is the most valuable and increasingly scarce skill in the AI era
Irreplaceable human qualities - Human qualities such as compassion, curiosity, and ingenuity cannot be replicated by AI, and remaining authentically human is a competitive advantage
Arg. 1Maricela Munoz argues that the qualities that make humans unique — compassion, curiosity, ingenuity, and critical thinking — cannot be replicated by any machine, however sophisticated. She contends that in a world increasingly saturated with AI, remaining authentically human and natural is itself a competitive advantage, particularly in professional settings where she evaluates candidates. She encourages young people to trust in their inherent human qualities rather than becoming dependent on AI.
She stated that 'our humanity and our curiosity will be always above any machine who can imitate, for example, critical thinking' , and noted that when hiring, she gets 'discouraged' when candidates 'depend too much on artificial intelligence', preferring 'people who remain very human and very natural and use technology as a tool' . She referenced singer John Legend, who said he rarely uses technology because 'he doesn't think that technology knows what love is, what pain is, and so many other human characteristics' .
on: Human qualities — curiosity, compassion, ingenuity, and the capacity for storytelling — remain irreplaceable by AI
Technology as human assistant - Technology should be used as an assistant to free humans from mundane tasks, allowing more time for reflection, vision, and out-of-the-box thinking
Arg. 2Maricela Munoz argues that the proper role of technology, including AI, is to serve as an assistant that handles routine and boring tasks, thereby freeing human beings to engage in higher-order activities such as reflection, imagination, and visionary thinking. Rather than replacing human creativity and wisdom, technology should create the conditions for more of it. She encourages people to use this freed time to be more visionary and to think beyond conventional boundaries.
She encouraged the audience to 'continue using technology to do our work better, to perhaps do less boring stuff and use the time to reflect and to take a pause and to imagine things out of the box, to be visionary' . She also affirmed that 'to create knowledge and to be wise and to be unique remains a quality and a project and a vision for a human being' .
on: AI is a tool to enhance human capability, not a replacement for human thinking and judgement
on: Whether AI is fundamentally ruining education or merely transforming it
Knowledge creation as human endeavour - The creation of knowledge, wisdom, and unique vision remains an essentially human project that no technology can fully replicate
Arg. 3Maricela Munoz argues that despite the impressive capabilities of AI, the creation of genuine knowledge, wisdom, and unique vision is and will remain an essentially human endeavour. She contends that there are dimensions of human experience — love, pain, and other emotional and existential qualities — that technology cannot understand or replicate, and that these form the foundation of truly meaningful knowledge creation. This makes the human project of learning and knowing irreplaceable.
She stated that 'to create knowledge and to be wise and to be unique remains a quality and a project and a vision for a human being' , and cited John Legend's view that technology cannot know 'what love is, what pain is, and so many other human characteristics' . She also affirmed that 'there are things that cannot be replicated even with the best of technologies, even with the technology that we still do not know today' .
on: Human qualities — curiosity, compassion, ingenuity, and the capacity for storytelling — remain irreplaceable by AI
Persistence of traditional schooling - Despite centuries of criticism and calls for reform, the traditional school format has remained remarkably unchanged even in the face of new technologies
Arg. 1Boris Engelson observes that the traditional school format has proven extraordinarily resistant to change, persisting despite centuries of criticism, pedagogical reform movements, and now the advent of AI. He notes that even discussions about AI and education tend to take place in the traditional school format — a panel or lecture — which he finds remarkable. He does not claim to fully understand why this is so, but suggests there must be deep structural reasons for the model's persistence.
He noted that 'from time immemorial, pupils hate school' and that 'one of the first texts of the history of writing dating back to Sumer was a complaint of a pupil' . He observed that he has 'heard during my life scores of discussion, analysis, with or without Piaget, with or without teachers, saying we should change the school system' , and that 'even here, listening to debates or lecture on AI, it is still the traditional school format' . He concluded that 'there must be deep reasons why this model cannot change' .
Paradox of knowledge vs. disruption - There is an inherent tension in society between the model of comprehensive established knowledge and the model of disruptive innovation, and both are necessary yet difficult to reconcile
Arg. 2Boris Engelson identifies a fundamental paradox in how societies approach knowledge: there is an inherent tension between the ideal of comprehensive, encyclopaedic knowledge (exemplified by Renaissance figures like Pico della Mirandola) and the ideal of disruptive, innovative thinking (exemplified by Einstein). Both models are necessary for society to function and progress, yet they are fundamentally in tension with each other. Common sense can support critical thinking but can also be its enemy, as truly disruptive ideas often challenge common sense entirely.
He contrasted Pico della Mirandola, the Renaissance man who 'knew everything', with Einstein, the hero of disruption and innovation, noting that 'these two models do not really fit together, and still we need both of them in our life' . He observed that 'common sense can be very useful for esprit critique, and at the same time, the theory of relativity was challenging any form of common sense' .
on: Whether common sense supports or undermines critical thinking
Session Knowledge Graph
Speakers · Topics · Arguments · Relationships
Catalin Radu explicitly stated that AI 'is indeed a tool to be used by human in a proper way to get better results, more effective, more productive' . Maricela Munoz encouraged the audience to 'continue using technology to do our work better, to perhaps do less boring stuff and use the time to reflect' . Hong Guan argued that 'AI shouldn't replace education. AI should push us to make education better' . Nabil Naoumi illustrated how AI rapidly produces drafts so that expert groups can focus on refinement and decision-making . All speakers consistently framed AI as an instrument in service of human goals rather than a substitute for human agency.
AI as professional tool - AI is a tool to be used by humans to achieve better, more effective, and more productive results, not a replacement for human vision and judgement
Technology as human assistant - Technology should be used as an assistant to free humans from mundane tasks, allowing more time for reflection, vision, and out-of-the-box thinking
AI as education enhancer - AI should not replace education but push universities to help students ask better questions and act with responsibility
AI for efficient drafting - AI saves significant time by producing drafts and compiling information rapidly, allowing expert groups to focus on refinement and decision-making
Ana Paula stated that 'critical thinking is coming at a premium now because information is now cheap' , and that 'you need to know what is correct, you need to know how to make a decision in a sea of information and contradictory information' . Hong Guan included 'critical judgment' as the fifth and arguably most important meta-capability, noting it is 'most important as other force, I think, nowadays' . Student 3 proposed that students must engage in critical reflection before using AI, not only after . Maricela Munoz affirmed that 'our humanity and our curiosity will be always above any machine who can imitate, for example, critical thinking' . The convergence across speakers from different backgrounds - professional, academic, and student - underscores the strength of this consensus.
Premium on critical thinking - Critical thinking is becoming increasingly precious and rare as information becomes cheap and ubiquitous through AI
Five meta-capabilities framework - Five meta-capabilities critical for the AI era: learning agility, execution capability, communication skills, leadership potential, and critical judgment
Pre-AI critical reflection - Before using AI, students should first clarify their own understanding and questions so they can critically evaluate the AI's response
Technology as a mirror for humanity - Technology acts as a mirror, prompting humanity to reflect on who we are, and we must accept our imperfections rather than seek technological optimisation
Irreplaceable human qualities - Human qualities such as compassion, curiosity, and ingenuity cannot be replicated by AI, and remaining authentically human is a competitive advantage
Student 2 described classmates who upload entire textbooks to AI tools the day before exams, 'pass the test with flying colors. But then the day after the exam, they forget everything' , raising the question 'what's the point of being in school if you're just going to do this?' . Student 3 observed that AI answers make students feel they understand, causing them to 'stop asking the further questions' , leaving them 'filling the knowledge, but cannot fully understand the meaning of it' . Jovan Kurbalija acknowledged the reality that students will use AI tools if available , while Hong Guan responded with the institutional solution of oral examinations to distinguish genuine learners from those who cheat .
AI enabling superficial learning - Students using AI tools to pass exams without genuinely learning the material, raising questions about the purpose of education
False sense of knowledge from AI - AI generates answers that make students feel satisfied, causing them to stop asking further questions and preventing deep understanding
Pedagogy must reflect reality - Pedagogy must align with common-sense human behaviour; if students have access to AI tools, they will naturally use them, making it futile to ignore this reality
Oral exams as AI-proof assessment - Oral examinations and project-based assessments can effectively distinguish genuine learners from those who rely on AI to cheat
Maricela Munoz argued that 'there are things that cannot be replicated even with the best of technologies' and cited John Legend's view that technology cannot know 'what love is, what pain is, and so many other human characteristics' . Jovan Kurbalija contended that storytelling is 'so powerful, so deep in the human nature, in our emotions, in who we are' , and that AI threatens this fundamental human capability . Hao Liu affirmed that the motivational power of storytelling 'you cannot take that from GPT, you cannot take that from DeepSeek' . Ana Paula stressed that 'to create knowledge and to be wise and to be unique remains a quality and a project and a vision for a human being' .
Irreplaceable human qualities - Human qualities such as compassion, curiosity, and ingenuity cannot be replicated by AI, and remaining authentically human is a competitive advantage
Knowledge creation as human endeavour - The creation of knowledge, wisdom, and unique vision remains an essentially human project that no technology can fully replicate
Two core learning methods: observing/apprenticeship and storytelling have existed since time immemorial - Learning by observing and narrating
Storytelling as natural motivation - Storytelling is a powerful motivational tool that drives people toward shared dreams without material rewards
Technology as a mirror for humanity - Technology acts as a mirror, prompting humanity to reflect on who we are, and we must accept our imperfections rather than seek technological optimisation
Ana Paula argued that 'adaptability is actually a skill. It can be learned. It can be improved. It's not an innate, it's not something that you were born with' , drawing on her own career trajectory from electronics engineering through multiple fields . Hong Guan placed 'learning agility' - knowing how to learn and adapt quickly - as the first and foundational meta-capability . Catalin Radu warned that organisations not yet implementing AI are 'already late' , implying that adaptability to technological change is now mandatory for survival . Nabil Naoumi drew a parallel with the arrival of the internet 25 years ago , illustrating that continuous adaptation to new technologies has always been necessary.
Adaptability as a learnable skill - Adaptability and continuous learning are learnable skills, not innate traits, and are essential for navigating an unpredictable future
Lifelong learning commitment - A commitment to lifelong learning, established early in life, is the most durable foundation for a career, as the specific knowledge studied often becomes obsolete
Five meta-capabilities framework - Five meta-capabilities critical for the AI era: learning agility, execution capability, communication skills, leadership potential, and critical judgment
Urgent organisational AI adoption - Organisations must implement AI system-wide and change their processes now, as those who have not started are already behind
Both Catalin Radu and Nabil Naoumi, as senior aviation and international organisation professionals, share the view that AI integration in professional contexts is not a future aspiration but a present imperative. Radu stated that 'it's already late' for organisations that have not begun AI implementation and that it is 'mandatory' for survival . Naoumi echoed this, stating 'if you just start now to think about AI, you are already too late. It is part of our daily business' . Both also emphasised AI's practical value in saving time and improving efficiency, with Naoumi describing how AI produces meeting drafts in seconds and Radu noting he no longer needs multiple assistants for administrative tasks . Student 2 and Jovan Kurbalija share a deep concern about the erosion of genuine intellectual engagement in universities. Student 2 observed that classmates pass exams using AI but 'forget everything' the next day and worried about 'who's going to be writing all these new research papers, who's going to be making new discoveries' . Kurbalija responded with equal concern, stating it was 'very worrying if at Stanford University you have this startup obsession, not people reflecting' , and affirming that university should be 'a place where you spend free time... to think, to reflect, to stay back, to have like-minded people, to engage, to question your thoughts' . Both see the current trajectory as a threat to the deeper purpose of higher education. Both Ana Paula and Hong Guan, approaching the question from practitioner and academic perspectives respectively, converge on the view that the most important educational outcomes in the AI era are meta-skills rather than specific knowledge. Hong Guan's framework places learning agility first — 'students must learn how to learn' — and culminates in critical judgment . Ana Paula independently arrived at the same conclusion, arguing that adaptability 'can be learned, it can be improved' and that 'continuous learning' is among 'the skills that will propel into the future' . Both also stress that AI should elevate rather than replace education, with Hong Guan stating 'AI should push us to make education better' and Ana Paula noting that 'when the teacher already expects you to use AI, oh, the bar is so high' . Maricela Munoz, Jovan Kurbalija, and Ana Paula all share a philosophical conviction that the AI era, rather than diminishing human value, actually foregrounds what is most distinctively human. Munoz argued that 'our humanity and our curiosity will be always above any machine' and that 'to create knowledge and to be wise and to be unique remains a quality and a project and a vision for a human being' . Kurbalija drew on Pope Francis's encyclical to argue that 'we have a right to be imperfect and in our imperfection is our ultimate purpose and beauty of the life' . Ana Paula grounded this philosophically in her personal experience, noting that a commitment to learning established at age 13 or 14 has sustained her through multiple technological revolutions, suggesting that human purpose and curiosity are the enduring constants. The three speakers from BIT's School of Global Governance share a coherent and complementary view of what practical capability looks like in the AI era. Student 1 argued that 'everyone can use AI to generate a lot of solutions, ideas... within five minutes' but the real question is 'how can we turn the AI-generated contents into real practice?' . Student 4 extended this by arguing that communication competence now includes giving precise instructions to AI systems, which 'didn't weaken the communication competence importance but also like strengthen it' . Hong Guan's framework encompasses both, placing execution capability second and communication skills third among the five meta-capabilities, and using 'real text and project-based learning' to develop them — exactly the approach Student 1 described experiencing .
It is somewhat unexpected that a self-described contrarian (Boris Engelson), a senior diplomat and educator (Jovan Kurbalija), and a first-year Stanford student (Student 2) would converge on a shared sense of unease about the persistence and inadequacy of traditional educational structures. Engelson observed that 'from time immemorial, pupils hate school' and that despite centuries of reform calls, 'the format remains nearly intact through the century and even with the most modern technologies' . Kurbalija, who might have been expected to champion institutional solutions, instead expressed worry that the old model is simply being perpetuated . Student 2, from the inside of one of the world's most prestigious universities, questioned 'what's the point of being in school if you're just going to do this?' . The consensus across such different vantage points - philosophical, institutional, and experiential - lends unusual weight to the concern.
Jovan Kurbalija explicitly acknowledged his uncertainty about whether Chinese traditions shared the same emphasis on apprenticeship and storytelling as Euro-Mediterranean traditions, stating 'I know very little about Chinese traditions on apprenticeship and storytelling. Is it the same?' . The response from the Chinese participants - Hong Guan and Hao Liu - confirmed that these traditions are indeed shared. Hao Liu affirmed that storytelling is 'critically important' throughout Chinese history and that 'all of the successful people are good storytelling people' . Hong Guan's framework, while framed in contemporary educational terms, implicitly reflects the same dual emphasis on practical execution (apprenticeship) and communication/storytelling. This cross-cultural consensus was unexpected given Kurbalija's stated uncertainty and represents a genuinely surprising finding of the session.
It is unexpected that a first-year Stanford student would articulate concerns that go beyond personal career anxiety to encompass systemic worries about knowledge generation for humanity. Student 2 stated explicitly that she was 'less so about my own employment, which is also very important, but more so about who's going to be writing all these new research papers, who's going to be making new discoveries in all of these different fields' . This systemic framing closely mirrors the concerns expressed by senior figures such as Kurbalija, who worried about the erosion of storytelling and reflective capacity , and Ana Paula, who noted that 'the future of work is still a burden' and that 'nobody knows where we are going' . The alignment between a young student's instinctive concerns and the considered views of experienced professionals represents a notable and unexpected convergence.
While the problem of AI-assisted cheating was widely acknowledged, it was somewhat unexpected that the discussion quickly converged on a concrete institutional solution - oral examination - rather than remaining at the level of concern. Hong Guan described BIT's use of oral examinations with multiple rounds of questioning , asserting that this model reliably distinguishes genuine learners from those who cheat . Jovan Kurbalija, who had framed the problem in broad pedagogical terms, immediately validated this approach by asking Student 2 whether she had actually read the 500-page textbook , implicitly endorsing the idea that direct questioning reveals genuine understanding. Hao Liu reinforced this by asking students to speak without manuscripts during the session itself , effectively demonstrating the principle in real time. The consensus around this practical solution, emerging spontaneously across speakers, was unexpected.
The discussion revealed a remarkably high degree of consensus across a diverse group of speakers - including senior international organisation professionals, Chinese and European academics, a Stanford undergraduate, and a self-described contrarian - on several core themes. First, there was near-universal agreement that AI is a powerful tool that must remain in service of human agency rather than replacing it, with every speaker framing AI instrumentally rather than as an autonomous force. Second, there was strong consensus that critical thinking, adaptability, and continuous learning are the most valuable skills in the AI era precisely because AI has made raw information cheap and abundant. Third, speakers converged on the view that AI poses a genuine threat to deep learning and academic integrity when students use it to bypass rather than enhance the learning process, and that institutional responses such as oral examinations and project-based assessment offer practical remedies. Fourth, there was broad agreement that distinctively human qualities - storytelling, compassion, curiosity, ingenuity, and the capacity for reflection - cannot be replicated by AI and represent humanity's enduring competitive advantage. Finally, a cross-cultural consensus emerged, somewhat unexpectedly, that Chinese and Euro-Mediterranean traditions share the same fundamental learning methods of apprenticeship and storytelling, suggesting that the challenges and responses to AI in education may be more universal than culturally specific. The one area of productive tension rather than consensus was between Boris Engelson's sceptical observation that educational systems have always resisted reform and may continue to do so , and the more optimistic institutional responses offered by Hong Guan and others - a tension that was acknowledged but not resolved within the session.
Student 2 argued forcefully that 'AI is ruining education' , describing classmates who upload entire textbooks to AI tools, pass exams 'with flying colors' but 'forget everything' the day after , and questioned 'what's the point of being in school if you're just going to do this?' . By contrast, Hong Guan stated clearly that 'AI shouldn't replace education. AI should push us to make education better' . Catalin Radu framed AI as 'a tool to be used by human in a proper way to get better results' , while Ana Paula argued that learning itself is evolving and 'the way that the topics are being taught is also evolving' . Maricela Munoz encouraged using technology 'to do our work better, to perhaps do less boring stuff and use the time to reflect' , presenting a fundamentally more optimistic view than Student 2's alarm.
AI enabling superficial learning - Students using AI tools to pass exams without genuinely learning the material, raising questions about the purpose of education
AI as education enhancer - AI should not replace education but push universities to help students ask better questions and act with responsibility
AI as professional tool - AI is a tool to be used by humans to achieve better, more effective, and more productive results, not a replacement for human vision and judgement
Adaptability as a learnable skill - Adaptability and continuous learning are learnable skills, not innate traits, and are essential for navigating an unpredictable future
Technology as human assistant - Technology should be used as an assistant to free humans from mundane tasks, allowing more time for reflection, vision, and out-of-the-box thinking
Jovan Kurbalija explicitly posed the question 'Is it a problem or not?' regarding the decline of storytelling skills, stating 'I feel it's a problem because by telling the story, you construct the narrative' , while also acknowledging 'maybe it's not a problem' . Hao Liu affirmed storytelling as critically important, arguing 'you cannot take that from GPT, you cannot take that from DeepSeek' . However, Catalin Radu's framing of AI as a productivity tool implicitly suggests that if AI can handle drafting and narrative tasks, this is a benefit rather than a loss, representing a different weighting of the value of human storytelling versus efficient output.
Two core learning methods: observing/apprenticeship and storytelling have existed since time immemorial - Learning by observing and narrating
Storytelling as natural motivation - Storytelling is a powerful motivational tool that drives people toward shared dreams without material rewards
AI as professional tool - AI is a tool to be used by humans to achieve better, more effective, and more productive results, not a replacement for human vision and judgement
Catalin Radu stated emphatically that 'it should happen now, today. So it's already late' and that AI adoption is 'mandatory, you know. You cannot survive the next era without starting to implement' . Nabil Naoumi echoed this, saying 'if you just start now to think about AI, you are already too late. It is part of our daily business' . Student 2, however, raised the counter-concern that widespread AI reliance raises the question of 'who's going to generate this new knowledge for the future of humanity?' , implying that the rush to adopt AI may be undermining the very knowledge base that AI depends upon .
Urgent organisational AI adoption - Organisations must implement AI system-wide and change their processes now, as those who have not started are already behind
AI for efficient drafting - AI saves significant time by producing drafts and compiling information rapidly, allowing expert groups to focus on refinement and decision-making
Threat to future knowledge generation - The concern that if everyone relies on AI, no one will generate new knowledge or make genuine academic discoveries
Hong Guan argued that BIT has effectively solved the AI cheating problem through oral examinations: 'We don't use that model of examination. We have oral examination. You don't have time to prepare to using AI' , asserting that 'those are people who are really reading that 500 pages. We definitely to be distinguished students' . Student 2's account of Stanford, however, painted a picture of an institution where written assessments remain the norm and AI-assisted cheating is widespread . Jovan Kurbalija noted this was 'not obviously... an example from Stanford, but it is all over the place. Because we are using the old model' , suggesting that most institutions, unlike BIT, have not adapted their assessment models.
Oral exams as AI-proof assessment - Oral examinations and project-based assessments can effectively distinguish genuine learners from those who rely on AI to cheat
AI enabling superficial learning - Students using AI tools to pass exams without genuinely learning the material, raising questions about the purpose of education
University culture and reflection - The startup obsession at elite universities risks crowding out genuine reflection and the deeper purpose of higher education as a space for thinking and questioning
Jovan Kurbalija argued that pedagogy must be grounded in 'common sense behavior of people' , using the example of a student who will naturally use AI to write an essay after a late night . He presented common sense as a practical guide for educational design. Boris Engelson, however, challenged this framing, noting that Kurbalija himself had taught him that 'common sense has one dangerous element that is a common and very often common sense is not critical thinking, it's just very often cacophony' . Engelson further argued that 'the theory of relativity was challenging any form of common sense' , suggesting that the most important intellectual breakthroughs actively contradict common sense, making it an unreliable foundation for critical thinking.
Pedagogy must reflect reality - Pedagogy must align with common-sense human behaviour; if students have access to AI tools, they will naturally use them, making it futile to ignore this reality
Paradox of knowledge vs. disruption - There is an inherent tension in society between the model of comprehensive established knowledge and the model of disruptive innovation, and both are necessary yet difficult to reconcile
It was unexpected that a student from Stanford University - widely regarded as one of the world's leading institutions - would present such a damning account of her own university's educational culture, describing widespread AI-assisted cheating and a 'startup obsession' that crowds out genuine intellectual reflection . Jovan Kurbalija validated this concern, calling it 'very worrying' and noting the problem is 'all over the place. Because we are using the old model' . Hong Guan's response was equally unexpected in its confidence: 'No, we don't have that hesitation because for us it's easy. We don't use that model of examination' , effectively positioning BIT as having already solved a problem that Stanford has not. This created an unexpected dynamic where a Chinese university was presented as more pedagogically adaptive than a leading American institution, inverting common assumptions about educational innovation.
Student 2 made a philosophically pointed argument that 'AI doesn't understand what it's saying' and 'is just a large language model that synthesizes everything that you put into it' , concluding that AI cannot generate truly new knowledge and that if everyone relies on it, 'who's going to generate this new knowledge for the future of humanity?' . This was unexpected because it directly challenged the optimistic framing of AI as a driver of innovation that pervaded the professional speakers' contributions. Catalin Radu and Nabil Naoumi, both senior professionals, enthusiastically endorsed AI adoption and pointed to AI's operational capabilities in aviation as evidence of its transformative power, without engaging with the epistemological concern that Student 2 raised about whether AI can produce genuinely new knowledge rather than recombining existing human knowledge.
Boris Engelson raised the unexpected observation that despite centuries of criticism - noting that 'one of the first texts of the history of writing dating back to Sumer was a complaint of a pupil' - the traditional school format has barely changed, and suggested 'there must be deep reasons why this model cannot change' , implying the persistence may reflect something structurally necessary rather than mere institutional inertia. This contrasted with Hong Guan's confident claim that BIT has successfully reformed its assessment model through oral exams , and with Jovan Kurbalija's call for pedagogy to adapt to common-sense behaviour . The unexpected element is Engelson's contrarian suggestion that the very reformers in the room were conducting their discussion about educational reform in 'the traditional school format' , implicitly questioning whether any of the proposed reforms would actually take hold.
Jovan Kurbalija drew on Pope Francis's encyclical to argue that 'we have a right to be imperfect and in our imperfection is our ultimate purpose and beauty of the life' , suggesting that the drive to optimise humans through technology is philosophically misguided. This created an unexpected tension with Catalin Radu's urgent call for comprehensive AI adoption, framed in the language of survival and competitive necessity: 'it's mandatory, you know. You cannot survive the next era without starting to implement' . Maricela Munoz partially bridged this gap by affirming human qualities as irreplaceable while also endorsing technology as an assistant , but the underlying tension between Kurbalija's philosophical acceptance of human imperfection and Radu's techno-optimistic drive for organisational optimisation was not explicitly resolved in the discussion.
The discussion revealed a moderate level of disagreement structured around several fault lines: (1) whether AI is fundamentally harmful or beneficial to education; (2) whether traditional educational institutions are capable of adapting to the AI era; (3) whether the urgency of AI adoption in organisations is as pressing as senior professionals claim; (4) whether human imperfection should be accepted or overcome; and (5) whether common sense is a reliable guide for pedagogy or a barrier to genuine critical thinking. The most substantive disagreements emerged between the student voices - particularly Student 2's account of AI undermining education at Stanford - and the professional voices of Catalin Radu and Nabil Naoumi, who enthusiastically endorsed rapid AI adoption . Hong Guan's confident claim that BIT has already solved the AI cheating problem through oral exams created an unexpected contrast with the picture painted of Stanford. Boris Engelson's contrarian observations about the persistence of traditional schooling and the paradox between encyclopaedic knowledge and disruptive innovation added philosophical depth to what might otherwise have been a consensus-oriented discussion.
All speakers agreed that critical thinking is essential in the AI era, but differed on how to cultivate it. Ana Paula argued it is 'coming at a premium now because information is now cheap' and that it is a learnable skill . Hong Guan included 'critical judgment' as the fifth meta-capability, describing it as 'most important' . Student 3 proposed a specific practical method: engaging in self-questioning before using AI so that 'we can have our maybe better judgment about the response' . Jovan Kurbalija framed the challenge as preserving storytelling as a vehicle for constructing narrative and critical thought . Maricela Munoz argued that human 'curiosity will be always above any machine who can imitate, for example, critical thinking' . While all agreed on the importance of critical thinking, they diverged on whether it should be cultivated through pre-AI reflection , oral assessment , apprenticeship , or simply trusting inherent human qualities .
Premium on critical thinking - Critical thinking is becoming increasingly precious and rare as information becomes cheap and ubiquitous through AI Five meta-capabilities framework - Five meta-capabilities critical for the AI era: learning agility, execution capability, communication skills, leadership potential, and critical judgment Pre-AI critical reflection - Before using AI, students should first clarify their own understanding and questions so they can critically evaluate the AI's response Two core learning methods: observing/apprenticeship and storytelling have existed since time immemorial - Learning by observing and narrating Irreplaceable human qualities - Human qualities such as compassion, curiosity, and ingenuity cannot be replicated by AI, and remaining authentically human is a competitive advantage
All speakers agreed that AI is, at least in principle, a tool rather than a replacement for human intelligence. Catalin Radu stated AI 'is indeed a tool to be used by human in a proper way' , and Maricela Munoz encouraged using it 'as an assistant' . Nabil Naoumi described AI as providing raw material that humans then work on . Ana Paula acknowledged AI helps get answers faster but noted 'the questions got much harder' . However, they disagreed on the degree to which this ideal is being realised in practice: Student 2 observed that in reality, students are using AI not as a tool for enhancement but as a substitute for learning , questioning whether the 'tool' framing reflects actual behaviour .
AI as professional tool - AI is a tool to be used by humans to achieve better, more effective, and more productive results, not a replacement for human vision and judgement AI for efficient drafting - AI saves significant time by producing drafts and compiling information rapidly, allowing expert groups to focus on refinement and decision-making Adaptability as a learnable skill - Adaptability and continuous learning are learnable skills, not innate traits, and are essential for navigating an unpredictable future Technology as human assistant - Technology should be used as an assistant to free humans from mundane tasks, allowing more time for reflection, vision, and out-of-the-box thinking Threat to future knowledge generation - The concern that if everyone relies on AI, no one will generate new knowledge or make genuine academic discoveries
All speakers agreed that educational systems need to change in the AI era, but differed significantly on how and how much. Jovan Kurbalija argued that pedagogy must align with common-sense behaviour and that the challenge is preserving storytelling skills . Hong Guan proposed a structured framework of five meta-capabilities and practical reforms such as oral exams [88-109, 188-192]. Ana Paula emphasised adaptability and lifelong learning as the core response . Boris Engelson, however, challenged the premise that change is achievable, observing that 'the format remains nearly intact through the century and even with the most modern technologies' , suggesting deep structural reasons prevent educational reform .
Pedagogy must reflect reality - Pedagogy must align with common-sense human behaviour; if students have access to AI tools, they will naturally use them, making it futile to ignore this reality AI as education enhancer - AI should not replace education but push universities to help students ask better questions and act with responsibility Lifelong learning commitment - A commitment to lifelong learning, established early in life, is the most durable foundation for a career, as the specific knowledge studied often becomes obsolete Persistence of traditional schooling - Despite centuries of criticism and calls for reform, the traditional school format has remained remarkably unchanged even in the face of new technologies
Student 1, Student 4, and Hong Guan all agreed that practical, applied skills are the critical differentiators in the AI era, but emphasised different dimensions. Student 1 focused on execution capability, arguing that 'how can we turn the AI-generated contents into real practice?' is the key question , and described project-based learning as the vehicle for developing this . Student 4 argued that communication competence has expanded to include human-AI interaction, with precise prompt-crafting now a core skill . Hong Guan's framework encompassed both, listing execution capability and communication skills among the five meta-capabilities, but also included learning agility, leadership potential, and critical judgment , suggesting a broader and more institutionally structured vision than the students' more experiential accounts.
Execution over idea generation - AI can generate ideas and solutions rapidly, but the real challenge is turning AI-generated content into real-world practice Expanded communication competence - Communication competence now extends beyond human-to-human interaction to include precise instruction-giving to AI systems Five meta-capabilities framework - Five meta-capabilities critical for the AI era: learning agility, execution capability, communication skills, leadership potential, and critical judgment
- Learning has fundamentally relied on two core methods since time immemorial: apprenticeship (learning by observing) and storytelling (learning by narrating and constructing narratives). AI now threatens the latter by making essay and story generation trivially easy.
- Five meta-capabilities are identified as critical for the AI era: learning agility, execution capability, communication skills, leadership potential, and critical judgment. These should form the basis of redesigned educational frameworks.
- AI should be treated as a tool that enhances education rather than replaces it. Universities must shift their role from providing answers to helping students ask better questions and exercise responsible judgement.
- The most pressing challenge AI poses to education is not access to information but the erosion of deep learning, critical thinking, and genuine knowledge construction. Students who use AI to pass exams without understanding the material retain nothing and contribute nothing new to human knowledge.
- Critical thinking is becoming increasingly rare and therefore increasingly valuable, precisely because information is now cheap and ubiquitous. The ability to evaluate, judge, and decide in a sea of contradictory information is a premium human skill.
- Adaptability and continuous learning are learnable skills, not innate traits. Given that the specific knowledge studied today is likely to become obsolete, a lifelong commitment to learning is the most durable professional foundation.
- Execution capability — the ability to translate AI-generated ideas into real-world results — is the key differentiator in the AI era, since idea generation itself has been largely commoditised by AI tools.
- Communication competence has expanded beyond human-to-human interaction to include the ability to give precise, effective instructions to AI systems, making it more important than ever.
- Organisations must implement AI system-wide and reform their processes urgently. Treating AI as the responsibility of a single designated person is insufficient; those who have not yet begun systemic AI integration are already behind.
- AI saves significant time by rapidly producing drafts and compiling information, freeing expert groups to focus on refinement, critical evaluation, and decision-making rather than information gathering.
- In professional domains such as aviation, AI is already deeply embedded in critical operational systems, including autonomous emergency landing capabilities, illustrating that AI integration in high-stakes fields is not a future prospect but a present reality.
- Human qualities — compassion, curiosity, ingenuity, and the capacity for genuine emotional understanding — cannot be replicated by AI. Remaining authentically human is both a moral imperative and a competitive advantage in hiring and professional contexts.
- Technology should serve as an assistant that frees humans from mundane tasks, creating more time for reflection, vision, and creative thinking rather than replacing human agency.
- Despite centuries of criticism and repeated calls for reform, the traditional school format has proven remarkably persistent even in the face of transformative technologies, suggesting deep structural or social reasons for its durability.
- There is an inherent and unresolved tension in society between the pursuit of comprehensive established knowledge and the drive for disruptive innovation; both are necessary yet difficult to reconcile within a single educational model.
- Pedagogy must align with actual human behaviour. Since students will naturally use available AI tools, educational systems that ignore this reality are ineffective. Assessment methods must evolve accordingly.
- The philosophical and humanistic traditions of diverse civilisations — Confucius, Ubuntu, Islamic thought, and Western philosophy — offer enduring wisdom relevant to navigating the ethical and existential questions raised by the AI era.
- Technology acts as a mirror that prompts humanity to reflect on its own nature and values. Accepting human imperfection, rather than seeking technological optimisation, is central to preserving human dignity and purpose.
- The creation of new knowledge, wisdom, and unique vision remains an essentially human endeavour. If all students rely on AI, the pipeline of genuinely new human-generated knowledge and discovery is at risk.
- A commitment to lifelong learning, ideally established early in life, is the most resilient foundation for a career, as the specific disciplines studied are likely to be transformed or rendered obsolete by technological change.
“My starting hypothesis is that we have two ways of learning: by observing (apprenticeship) and by hearing/telling stories. Now, storytelling in its last form — writing essays — is more or less done, because platforms can generate nice stories. If pedagogy is not associated with common sense behaviour of people, it's not good. Common sense behaviour is that you use tools that you have.”
“AI is ruining education. My classmates would upload the entire textbook into Claude or ChatGPT, cross-reference it the day before lecture, pass the test with flying colors, and then forget everything the day after the exam. If every person at Stanford is using AI, then who is going to generate new knowledge for the future of humanity?”
“We don't use that model of examination. We have oral examination. You don't have time to prepare using AI. We ask questions one round, a second round — you feel. We will evaluate your performance based on how you feel. We know how to torture you. We know how to challenge you.”
“Critical thinking is coming at a premium now because information is now cheap. The notion of holding expert power through information has already changed. Adaptability is actually a skill — it can be learned, it can be improved. It is not innate.”
“From time immemorial, pupils hate school. One of the first texts in the history of writing, dating back to Sumer, was a complaint of a pupil. I have heard scores of discussions saying we should change the school system. Today, even teachers and ministers hate school. And still, when we come here to listen to debates on AI, it is still the traditional school format. There must be deep reasons why this model cannot change.”
“We have a right to be imperfect. The Pope's encyclical Laudate Deum — or rather Dilexit Nos — argues that we should be accepted as imperfect as we are. We should not be optimised. In our imperfection is our ultimate purpose and the beauty of life. Technology is a commodity, but one useful thing it brings us is that it puts a mirror in front of us to see who we are.”
“Before using AI, we can ask ourselves: what exactly are we confused about? When we know that, we can have a better judgment about the response and have critical thoughts. We have to make critical practice longer — not only after using AI, but before using it.”
Is the loss of storytelling skills a genuine problem in the AI era, or is it irrelevant if AI can generate stories for us?
Kurbalija raised this as a central unresolved question, noting that storytelling is deeply embedded in human nature and cognition, but acknowledging that AI tools like DeepSeek and ChatGPT can now generate compelling narratives. Whether humans still need to develop this skill independently remains open for debate and further investigation.
How do Chinese educational and cultural traditions approach apprenticeship and storytelling as modes of learning, and how do they compare to Euro-Mediterranean traditions?
Kurbalija explicitly stated he knows very little about Chinese traditions on apprenticeship and storytelling and invited comparison. A deeper cross-cultural study of learning traditions could inform more globally inclusive pedagogical models for the AI era.
If students at elite universities like Stanford are using AI to pass exams without genuinely learning, who will generate new knowledge for the future of humanity?
This question strikes at the heart of the sustainability of academic knowledge production. If AI is used to shortcut learning rather than deepen it, the pipeline of original researchers, scientists, and thinkers may be at risk, warranting urgent investigation into the long-term effects of AI dependency in higher education.
What is the appropriate role of AI in higher education, and how should universities redesign assessment models to ensure genuine learning rather than AI-assisted performance?
Multiple participants highlighted the inadequacy of current assessment models in the face of AI tools. Hong Guan pointed to oral examinations as one solution, while others raised the broader question of how institutions should restructure evaluation to distinguish authentic learning from AI-generated outputs.
Should universities or institutions introduce differentiated degrees that distinguish between AI-assisted and non-AI-assisted work?
Hao Liu proposed this as a potential solution to the credentialing problem created by AI use in education. This idea warrants further research into its feasibility, ethical implications, and how it might be implemented across different educational systems and cultures.
Why has the traditional school format persisted largely unchanged across centuries, even in the face of major technological disruptions, and will AI finally change it?
Engelson observed that despite widespread dissatisfaction with schooling and repeated calls for reform, the traditional format endures. Understanding the deep structural, social, and psychological reasons for this persistence is essential before designing AI-era educational reforms.
How can societies simultaneously cultivate comprehensive established knowledge and disruptive innovative thinking, given that these two models are in tension with each other?
Engelson identified this as a fundamental paradox: the Renaissance ideal of total knowledge versus the Einsteinian ideal of radical disruption. Resolving or managing this tension is critical for designing educational systems that produce both well-rounded graduates and transformative innovators.
How can critical thinking be taught and assessed as a skill in an environment where AI provides instant, authoritative-seeming answers that discourage deeper questioning?
Student 3 observed that AI responses often give students a false sense of understanding, halting further inquiry. Ana Paula noted that critical thinking is becoming increasingly rare and valuable. Research is needed into pedagogical methods that cultivate genuine critical judgement before, during, and after AI use.
How is the concept of communication competence evolving in the AI era to include human-to-AI interaction, and how should this be incorporated into education?
Student 4 raised the point that effective communication now includes crafting precise prompts for AI systems. This suggests a need for research into how curricula should be updated to teach prompt engineering and human-AI interaction as core communication skills alongside traditional interpersonal communication.
To what extent is adaptability a learnable skill rather than an innate trait, and how should it be systematically taught in educational institutions?
Ana Paula asserted that adaptability and resilience are skills that can be developed, not just innate characteristics. Further research is needed into evidence-based methods for teaching these meta-skills, particularly given the rapid pace of technological and professional change.
What are the implications of AI and autonomous systems replacing human roles in safety-critical industries such as aviation, and how should education prepare students for this transition?
Naoumi described how AI is already capable of landing aircraft autonomously and that decision-makers are considering replacing pilots. This raises urgent questions about workforce preparation, ethical governance of autonomous systems, and how educational institutions should respond to such profound occupational disruption.
How should organisations implement AI at a systemic level rather than in isolated roles, and what change management frameworks are needed to support this?
Radu argued that organisations cannot simply designate one 'AI person' but must transform entire processes and systems. Research into organisational change management, AI integration strategies, and leadership development for the AI era is needed to support this transition effectively.
How can the philosophical and humanistic traditions of different cultures — including Confucianism, Buddhism, African Ubuntu, and Western philosophy — inform the governance and ethical use of AI?
Kurbalija referenced the Pope's encyclical Laudate Deum and various philosophical traditions as sources of wisdom for the AI era. Further interdisciplinary research into how diverse cultural and philosophical frameworks can contribute to AI ethics and governance would enrich the global conversation beyond Western-centric perspectives.
What does it mean to preserve human imperfection and dignity in an era where technology increasingly seeks to optimise human performance, and how should this principle shape AI policy?
Drawing on the papal encyclical, Kurbalija argued for a 'human right to be imperfect' as a counterweight to Silicon Valley's optimisation agenda. This raises important questions for AI policy, ethics, and education about the values that should underpin technological development and deployment.
How can experiential and project-based learning models, such as international study tours and apprenticeship programmes, be scaled and integrated into mainstream education to develop execution capability in the AI era?
Student 1 described the value of structured international practical learning experiences, and Hong Guan referenced project-based learning as a key pedagogical approach. Research is needed into how such models can be made accessible beyond elite institutions and how they can be systematically evaluated for effectiveness.
What is the long-term impact of AI dependency on students' ability to generate original ideas, and how can this be measured and mitigated?
The Stanford student expressed concern that widespread AI use is preventing genuine intellectual development, while Kurbalija noted the worrying prevalence of a startup-obsessed culture over reflective thinking. Longitudinal research into the cognitive and creative effects of AI dependency on students is urgently needed.
