https://www.unesco.org/en/articles/ai-and-future-education-disruptions-d...
The history of education is a history of disruption. Five thousand years ago, writing transformed memory into text. Six hundred years ago, Gutenberg’s printing press democratized knowledge. Three years ago, ChatGPT put generative AI in the hands of hundreds of millions.
For the first time in history, technology is not silent. It speaks back.
For generations, the image of the classroom looked the same: a teacher in front of a chalkboard, students sitting in rows, reading from textbooks. One of humanity’s oldest rituals – the slow passing of knowledge from one person to another.
That ritual is now being rewritten. Not by a new curriculum. Not by a global treaty. But by a new actor. One that never sleeps, never breathes, never stops ingesting information.
Artificial intelligence is the latest in a series of technological innovations that has changed how we learn, work, and communicate.
It is now altering the way knowledge is accessed and created, and already more than a billion people use AI chatbots regularly.
In education, this forces us to ask fundamental questions.
Do we still need classrooms, teachers, textbooks? If an AI system can write a PhD-level thesis in minutes and win international math competitions, what does it mean to be a student?
That is the magnitude of the disruption before us.
At UNESCO, we are reflecting on three dimensions of this transformation: the disruptions AI is creating in education, the dilemmas these disruptions raise, and the directions we must chart together.
These themes are explored in depth in UNESCO’s recent book, AI and the Future of Education: Disruptions, Dilemmas and Directions, which gathers global evidence, case studies, and policy recommendations.
Disruptions: how AI is reshaping classrooms
We hear a lot of rhetoric about how AI is disrupting education. What do these disruptions look like in practice?
First, AI as tutor. Chat-based assistants and adaptive platforms now provide explanations, practice, and feedback, often in local languages and even through low-bandwidth channels. In some contexts, this is the first time a learner has had access to one-on-one support.
Studies show that, when designed to encourage reasoning, these tools can improve outcomes. But when they simply hand out answers, students may learn less, not more. The difference lies in how the technology is used and supported.
Second, AI as co-professor. Across universities, it is already being used for research, writing, coding feedback, and assessment.
A recent survey of the UNESCO Chairs network found that nine in ten respondents use AI in their professional work, most often for writing and synthesis. At the same time, many expressed uncertainty about its pedagogical value and its implications for integrity. The opportunity is to free time for higher-value teaching. The risk is a drift toward automation.
Third, AI as companion. Many young people now turn to chatbots for advice, support, or late-night conversations.
Surveys in high-income countries suggest that a majority of teenagers have already experimented with AI companions, and some use them regularly for serious conversations. There is evidence that such tools can reduce stress and anxiety, but they also raise concerns about privacy, dependency, and the substitution of human relationships.
Across these roles—tutor, co-professor, and companion—AI can expand access and support, but it can also deepen inequality, flatten learning, and erode trust if not guided carefully.
Whether this disruption strengthens education or undermines it will depend on the choices we make.
Dilemmas: the hard choices being faced
Every disruption brings dilemmas.
- Should a school principal allow AI in classrooms – or restrict it to protect students’ cognitive development?
- Should a minister of education invest in AI platforms that could reach millions – or in hiring the 44 million additional teachers the world still needs?
- Should a head of government approve a new AI data centre that promises jobs – or consider the consumption of massive amounts of water and electricity?
None of these are simple yes-or-no choices. These demand foresight and public debate.
Directions: charting a human-centered path
But dilemmas are not dead ends. They are signposts to new directions.
At UNESCO’s Digital Learning Week, ministers affirmed a shared vision that AI in education must be human-centered, equitable, safe, and ethical.
From that dialogue, five priorities emerged.
First: confronting the AI divide. The digital divide is rapidly becoming an AI divide. Without electricity, meaningful connectivity and affordable devices, millions will be left behind. That is why sustainable digital infrastructure must be our first priority.
Second: safety and ethics are non-negotiable. Human rights and dignity must come first. We need clear frameworks so that AI tools in schools are transparent, accountable, and free from bias. This is why UNESCO adopted the world’s first global Recommendation on the Ethics of Artificial Intelligence, and why we issued Guidance on generative AI in education and research.
Third: teachers are irreplaceable. As the Santiago Consensus adopted at the World Summit on Teachers underlined: AI must support, not replace, educators. This means investing in professional development and skills. This is the principle behind UNESCO’s AI Competency Frameworks for teachers and students, already translated into multiple languages.
Fourth: localization and cultural relevance. AI must work for all, and all means all. AI must reflect the diversity of our societies, languages, and communities. That means building and adapting models locally and training them on culturally diverse and representative datasets.
Fifth: global solidarity and shared standards. No country can navigate this transition alone. International cooperation is essential to set standards, develop curricula, and exchange knowledge on emerging pedagogies.
Moving beyond binary thinking
In debates on AI and education, we often hear extremes.
Either AI will sharpen skills or dull them. Either it will replace teachers or empower them. Either it will save education or destroy it.
The truth lies somewhere in between.
And like most big disruptions, the impact of AI on education will be complex, contradictory, and unfolding all at once.
It will be determined in parliamentary debates and corporate labs, but also in classrooms, villages, and communities.
Depending on the choices we make today, the future of AI in education will be marked by expanded opportunity or by deepened inequality.
What we need – and what UNESCO has worked for 80 years to promote – is imagination, courage, ambition, and above all, collective action.
