Article content
Throughout history, access to knowledge has been limited. Education was largely reserved for those with access to scarce learning materials, and this shaped our understanding of what it means to be a smart person. Technology has changed this reality. Today, all of the world’s information is available in our pockets, and AI has proven to be a useful tool for explaining a wide range of topics.
Nowadays, people increasingly rely on machines to provide facts, explain historical events, or clarify complex subjects. If these capabilities are accessible to most people, then what is the definition of an educated person?
Certain professions, such as medicine, will always require large amounts of knowledge to be instantly available from memory. Yet overall, I believe we will increasingly see memorised information lose some of its value as the primary measure of intelligence.
What is the next metric we should be looking for?
In an age of information overload, it is not enough to find information quickly. We must also be able to determine which information is factually correct and useful, and which information should be ignored. I believe a person’s capabilities will be measured by their ability to reason, ask the right questions, make deductions, and connect different pieces of information.
Let us look at Estonia as an interesting case study. Estonia is culturally unique in many ways. As a small nation that has spent much of its history being governed by others, Estonians have developed a strong belief that education is one of the most valuable legacies children can inherit. No matter what happens, a person with knowledge and the ability to think can find a way forward.
This can also be seen in Estonian folklore. Many traditional stories revolve around outsmarting an opponent rather than overpowering them. We also find tales of small, clever creatures that use intelligence and ingenuity to achieve their goals, rather than stories centered on strength and valor.
I believe these cultural traits provide a strong foundation for developing the skills that are becoming increasingly valuable in today’s AI-driven world. These strengths are also reflected in PISA test results, where Estonia ranks first in Europe and performs exceptionally well on a global scale.
It is already noteworthy that PISA is beginning to measure students’ digital skills and habits. When we look at the results of the 2025 PISA tests, an interesting pattern emerges. On average across OECD countries, 46% of students use AI chatbots in their studies at least once a week. In Estonia, that number reaches 58%. The results also show that 63% of students across OECD countries learn at school how to verify information provided by AI tools, while in Estonia the figure stands at 73%.
We can clearly see that Estonian students are active users of AI solutions. At the same time, PISA results show that students are beginning to struggle with some of the skills that are most important in the age of AI: critically assessing information, connecting information from different sources, and thinking deeply about what they have read.
This is what the current AI Leap is all about in Estonia. It is an attempt to create a systematic change that supports students in developing the new skills they need in a world shaped by information overload and AI assistants.
Estonia has now reached a point where its AI Leap initiative is beginning to provide the first feedback on how AI in education is shaping the way we approach knowledge.
The report on the pilot year highlights several notable shifts in teaching practices and school culture. Among participating teachers, 64.9% reported moving away from controlling students’ use of artificial intelligence toward guiding and supporting its use, while 44% felt their role had evolved from being a primary source of knowledge to acting as a critical mediator.
At the same time, collaborative learning practices among teachers became more common, increasing from 59.7% to 67.1%. The growth was even more pronounced in AI-focused learning circles, which rose from 70.5% to 83.4%, suggesting that peer learning and the exchange of experiences among colleagues became one of the program’s most significant outcomes.
The report also points to changes in everyday pedagogical practice. The strongest growth was seen in the modification of homework assignments, which increased from 65.6% to 70.4%. Teachers also reported redesigning assessment systems, with this figure increasing from 28.5% to 30.7%. This latter finding is particularly noteworthy, as assessment practices are typically among the most difficult and slowest aspects of education to change.
Teachers provided a range of examples illustrating how artificial intelligence is already being integrated into the classroom. As one teacher explained, “I had ChatGPT simulate an athlete’s body during a competition while teaching hormonal balance.” Another described how AI has transformed lesson preparation: “This year I am experimenting with dramatising a novel using ChatGPT. The work saves hundreds of hours.” A third teacher highlighted AI’s role in encouraging critical thinking through source comparison: “Artificial intelligence impersonating a historical figure vs. the portrayal of a historical figure in a book text.”
Teachers will increasingly act as role models who demonstrate reflection, self-assessment, emotional regulation, goal-setting, and the ability to work toward those goals. In a situation where students often have more experience with AI than their teachers, this presents a significant challenge to the traditional hierarchy of the classroom, which needs to shift more toward a partnership.
More about the AI Leap.
