AI and the Future of Education: Disruptions, Dilemmas, and Directions
Presented and Translated by the AI Studies Unit

The original book consists of 165 pages and was translated from English.

The emergence of artificial intelligence (AI), particularly generative models, marks a pivotal moment in the history of education, comparable in impact to the invention of the printing press or the rise of the internet. AI is not merely a new technological tool; it is a transformative force reshaping concepts of learning, teaching, knowledge, and educational relationships themselves. UNESCO’s comprehensive report, “AI and the Future of Education: Disruptions, Dilemmas, and Directions,” explores these transformations through a collection of essays offering multiple, intersecting, and sometimes conflicting perspectives on AI’s impact on global education.

The Core Dilemma: Promises of Transformation vs. Risks of Deepening Inequalities

The report highlights the fundamental tension accompanying AI’s spread in education. While AI offers enormous potential for personalized learning, expanding access to knowledge, and freeing teachers from routine tasks, it carries significant risks. Over a third of the world’s population remains offline, meaning access to the latest AI models is limited to those with the necessary infrastructure, linguistic, and financial privileges. This gap determines not only who can use technology but also which knowledge, values, and languages dominate these systems, threatening to entrench Western-centric knowledge and marginalize local knowledge and languages. AI thus becomes a new arena for the struggle between inclusion and exclusion, commercial interests, and human-centered educational goals.

Philosophical Insights: Rethinking the Foundations of Education

The first section delves into philosophical questions raised by AI. In a conversation with philosopher Bayou Akomolafe, modern notions of the independent, rational “self” are challenged. Akomolafe views AI not merely as a tool or threat but as a “more-than-human interlocutor,” unsettling the ontological and epistemological certainties underpinning modern education. He advocates for adopting outdated educational models and “hosting” disruption and breakdowns as spaces for new knowledge generation, suggesting a shift from “curriculum as structure” to “curriculum as living assemblage,” where learning is relational, emergent, and responsive to non-human agencies, including AI.

In a complementary essay, Bing Song draws on Asian philosophical traditions, particularly Confucianism and Taoism, emphasizing harmony, relationality, and inner balance. She critiques the narrow “intelligence paradigm” of AI, which replicates aspects of human intelligence while neglecting wisdom, self-reflection, and ethical dimensions. She calls for integrating a “pedagogy of wisdom” into curricula to reflect on the nature of humans, self, mind, and reality, countering the risk of “automating humanity” according to machine standards.

Using a strong metaphor, Mary Rice and Joaquin Arguello compare AI’s integration into education to water. Just as water was the first teacher (through patience and gradual transformation), the first technology (through play and creativity), and the first scarce resource (due to exploitation), AI carries multiple faces. The essay warns against excessive resource consumption (water and energy) to run AI data centers and against transforming education into a detached, abstract experience, which risks reproducing patterns of colonialism and inequality.

The Power-Risk Dialectic: Cautious Adoption vs. Radical Rejection

Thinkers differ on the degree to which AI should be embraced in education. From an industrial perspective, Andreas Horn offers a practical vision, advocating for clear strategies that place “pedagogy first.” He emphasizes leveraging teachers’ expertise, making digital literacy and AI knowledge fundamental, enforcing strict ethical guidelines, and preparing students to lead an AI-rich world, not merely adapt to it.

In contrast, Emily M. Bender delivers a sharp critique, describing large language models as a “parlour trick” that mimics linguistic form without understanding or intent. She argues that any perceived intelligence is a projection of human cognition. She warns that these systems produce “personalized misinformation” for each student, reduce education to isolated knowledge accumulation, and undermine fundamental teacher-student relationships. Adopting such technologies in resource-limited systems risks wasting public funding and turning education from a public good into a corporate market.

Marcus Deyman and Robert Faro expand the discussion to “social imaginaries” of AI, ranging from “technological-utopian” views seeing AI as an educational savior, to “libertarian-cyber dystopias” promoting privatized education free from democratic oversight, and to “ecological disaster” scenarios due to massive energy and data consumption. They stress the active participation of educators in shaping these imaginaries, grounded in values of justice, inclusion, sustainability, and care.

Pedagogy and Assessment in the Age of AI

The next section examines AI’s direct impact on educational practice. Abiba Berhan, drawing on Paulo Freire’s critical pedagogy and embodied cognitive science, warns that education is inherently relational, dynamic, and ethical, and cannot be reduced to probabilistic patterns.

She notes that AI systems trained on past data tend to flatten human complexity and reproduce biases and systemic inequalities. She calls for resisting uncritical adoption of AI in classrooms until independent oversight and meaningful community participation mechanisms are in place.

Carla Irts and Paul Brinslow address the issue of “hyper-personalized” learning, which, while promising optimal learning, can isolate learners in “knowledge echo chambers,” undermine autonomy, and marginalize teachers. They propose a model where AI serves as a supportive third presence in collective intelligence, enhancing collaboration, empathy, and student agency.

In assessment, Mike Perkins and Jasper Roe offer a somewhat pessimistic view, arguing that generative AI signals the “end of assessment as we know it.” Traditional tests will no longer reliably measure real learning, deepening global inequities. They propose a tiered framework (e.g., AI Assessment Scale, AIAS) to guide educators in determining when AI supports or undermines learning.

Conversely, Bill Koop, Mary Kalantzis, and Akash Kumar Saini offer a more optimistic, forward-looking vision. They critique multiple-choice and standardized tests as outdated measures of “surface learning.” Instead, they envision “socio-cyber learning environments” where AI provides continuous, rich, formative feedback based on teacher-designed standards, reliable knowledge, and student’s zone of proximal development. Summative assessment becomes retrospective, and education refocuses on “complex cognitive performance” and uniquely human skills.

Re-centering the Human Teacher

Ching Sing Chai and colleagues draw on Martin Buber’s I-Thou philosophy and Gert Biesta’s educational goals (qualification, socialization, subjectification), emphasizing that education fundamentally depends on authentic human encounters. AI can effectively support “I-It” (instrumental) interactions but cannot participate in genuine “I-Thou” relationships. Teachers provide care, ethical challenge, and witness student growth—dimensions machines cannot replicate. They advocate for “triadic” collaborative models (I-Thou-It) in which AI complements rather than replaces human relationships.

Arfa Karimi translates these principles into a practical roadmap, proposing “seven transformations” for AI systems focused on care. These include co-design with teachers and students, trust and wellbeing reviews, transparent decision explanations, and teacher oversight of data, repositioning AI as a collaborative partner in a pedagogical ecosystem that preserves dignity and belonging.

Governance and Ethics: Toward Inclusive Justice

As AI penetrates education, governance and ethical issues intensify. Kaska Boraiska-Bumsta and Isak Enti Asari propose a “care ethics of design,” emphasizing that ethical considerations should be integrated from the outset through participatory design processes involving all stakeholders, especially the marginalized.

Calervo N. Gulson and Sam Sillar analyze the rise of “synthetic governance,” where educational decisions increasingly rely on algorithmic logic and data-driven systems. They caution against assuming these systems are neutral and call for critical democratic responses that expose power relations and reaffirm education as a public good.

Addressing Encoded Inequalities: Perspectives from the Global South and Marginalized Communities

Several essays provide alternative visions emphasizing justice and equity. Fokusi Marivan et al. propose AI integration in African higher education rooted locally, respecting cultural and linguistic diversity, prioritizing human agency and pedagogical care. AI systems should not only translate but adapt to diverse communicative styles and support marginalized languages.

Kiran Bhatia and Pail Aurora challenge paternalistic models framing young women in the Global South as “risks.” They advocate centering joy, creativity, and transformative agency, enabling young women to participate in shaping their digital futures and transform unequal power structures.

Yuchen Wang emphasizes conceptual clarity in inclusion, linking it to belonging, relationality, and collective learning rather than narrow personalization. Marlos Williams highlights the complexities faced by deaf or hard-of-hearing learners in resource-limited contexts, calling for co-designed, multi-modal AI systems complemented by human support, since “equity cannot be automated.”

AI Policy in Education: Geopolitics and Collective Meaning-Making

The report concludes with two policy perspectives. George Siemens offers a geopolitical analysis, noting AI as a state tool invested in strategically like military or economic power. He calls for educational systems leveraging AI while safeguarding human wellbeing.

Elka Tuomi reframes educational policy as “collective meaning-making” and “evolutionary experimentation” rather than linear implementation. She critiques the commodification of knowledge in the generative AI era, emphasizing human agency, social purpose, and capacity development as central educational goals.

Conclusion: Toward a Shared Future for AI in Education

AI use in education requires more than technical policy; it demands deeper dialogue, collective reflection, and ethical imagination. The narratives of AI and the future of education are still being written, and we bear collective responsibility to shape them with care, clarity, and courage. Whether through pedagogical design, political reform, or ethical governance, we must think, converse, and learn together to create an AI-enabled educational future that is inclusive, ethical, human-centered, and ecologically sustainable.

The goal is not unrestricted adoption nor total rejection of AI but to tame and guide it to serve the broader human vision of education as a fundamental right and global public good.