Can artificial intelligence make kids smarter?
AI in education could offer kids a chance to tap into higher learning skills, if schools are smart about its use, argues Kristina Kallas, Estonia’s Minister of Education.
When kids started using AI to learn, fears about the future of education became omnipresent. Teachers and parents panicked that a thinking machine could replace human educators, and students could offload their brain to technology and stop interacting with classmates. This could result in a generation who are innumerate, illiterate, antisocial and unskilled, as all they need are basic skills to prompt bots to complete any assignment. Even the kids realised learning could now be hacked. In a UK study from 2025, one in four students agreed that AI “makes it too easy for me to find the answers without doing the work myself.”
However Katrina Kallas, Minister of Education for Estonia since 2024, hopes that a controlled and responsible AI policy in schools can flip these concerns.
Often regarded as the most digitally advanced country in the world, Estonia is also ranked first in Europe according to PISA results.
She believes AI in school will mean more teachers, not fewer. Students could spend more time offline, not online. They could also be working in groups more often, not less. Above all, she hopes AI can help kids reach higher realms of thinking.
“We must make students think deeper and faster at an earlier age in education,” she says in an exclusive interview with Scoala9.
In other words, using AI wisely is a chance to make kids wiser.
“AI is here to push us to a higher order of thinking,” she says. “This is an evolutionary challenge we have to face. If we don’t do anything about AI, our knowledge will fall further down, and kids will be offloading it all to AI. If we don’t do it, computers will take over.“
50-year-old Kallas is co-leader of the Liberal Eesti 200 party, currently a partner in the Baltic state’s Government. A doctor in political science, she was an OSCE consultant, including for the Moldovan government on ethnic policy, and is current President of the Estonian Handball Association. At the heart of her mandate has been a program to build a public education system that can shape AI use in a way “that supports learning, equity and critical thinking” according to the Ministry. Her country of 1.38 million last year piloted a more proactive AI policy for post-16 education. Results will soon be coming in.
“Very few countries are attempting this at a system level,” she says, “which is why the outcomes are being closely watched.”
Fear number one: AI makes kids stupid
Firstly, there is the fear that AI is making students less intelligent. Studies argue that AI is impacting the long-term acquisition of skills and people’s ability to think critically. Using simple prompts, ChatGPT, Gemini or Grok takes only a few seconds to do kids’ homework, write their essays, solve their moral dilemmas and build visual, audio or video projects with little human contribution.
“The biggest concern is there will be a lot of cognitive offloading,” says Kallas, where the AI bot does all the work. A more positive approach is to use what she calls ‘cognitive outsourcing’: “With cognitive outsourcing, some of the cognitive capacities like memorising or analysing can be outsourced to AI, which can do it faster and more efficiently.”
The argument against AI making us stupider is the calculator analogy. Electronic calculators were introduced in the 1960s and became widely used by the 1970s in schools.
“When the calculators were invented we also did cognitive outsourcing to the calculators,” says the Minister. “The discussion [at that time] was that people will become dumber and no one will learn their multiplication tables by heart. People will be idiots. What actually happened is that by grade 12 today, kids graduate high school learning at a higher level than in the 1960s, because of calculators. This pushed students to be faster and more efficient in maths.
“The same with Google search. The talk at the time [the 2000s] was that students will be stupid and won’t memorise. That didn’t happen. They became much more skilled in analysing information very fast.”
She argues that the education system can withstand any major technological development that speeds up our information gathering “as long as we are capable of analysing what we outsourced and what kind of results the outsourcing produces.”
“For that we need higher order thinking skills,” she says.
Kallas uses Benjamin Bloom’s 1950s taxonomy of learning objectives as a framework for how to understand education in the age of AI.
At the bottom are lower-order thinking skills, like remembering, understanding and applying. These are the repetitive acts like learning to ride a bike, swimming, skating and languages.
The higher order are analyzing, evaluating and creating, such as writing poetry or making ethical decisions. This order generates new knowledge. These skills needed to be accessed in post-16 education, with the help of AI.
“Training the teachers to think that lower order thinking skills are not just enough,” she says. “We need to do more analysis work, more creative thinking and critical work.”
Estonia: at the edtech frontline
Estonia has been at the frontline of digital development, especially when it comes to integrating technology with public services. As a country with just over one million people and a small geographical area, it has arguably been the right location to act as a sandbox for state-wide technological revolutions.
In education, in the mid-1990s, Estonia’s ‘Tiger Leap’ programme brought computers and the internet to every school, at a time when only 0.5 per cent of the world population was online.
Since the AI revolution in 2022, up to 90 percent of Estonian students in grades six to 12 were using AI for school, mostly to speed up tasks rather than build understanding. This development lacked guidance. At the same time, only 53 percent of teachers were using AI, according to a 2024 survey, revealing a mismatch between student and teacher experience.
So last year Estonia began its AI Leap (TI‐Hüpe), for all upper-secondary school students. A pilot scheme reached almost all the country’s 156 high schools, involving 4,900 teachers and 20,000 students.
All high school teachers have taken two day online training courses. This means inviting Silicon Valley into the classroom, where AI Leap collaborates with Google’s Gemini and OpenAI’s ChatGPT on AI applications. Kallas argues that partnering with the private sector brings “innovation into the system.”
This four-million-euro scheme was financed 50 percent by the Ministry of Education and Research, and 50 percent from companies, philanthropy funds and entrepreneurs.
Teachers in Estonia have weekly or biweekly meetings at “hubs” where they aim to reflect and learn from each other on how to integrate AI into teaching.
“The biggest part of the teaching job is self-reflection,” says Kallas.
“AI cannot be a random tool that some kids use”
But what is happening on the ground, at school and at home? A problem with some schools in Europe is that kids are using AI to complete homework assignments, while others don’t. Yet both may end up with the same marks. Kids feel like losers for not using AI short-cuts.
“This is what happens when the teacher is not in control of the learning process and when teachers do not integrate AI into the process,” says Kallas. “This allows AI to be a random tool that some kids use and some kids do not. Teachers need to be aware that AI is there and is in everyone’s pocket and the way you design the learning process you have to take that into consideration.”
Rather than ban the practice or pretend AI isn’t there, teachers need to redesign projects with AI in mind. They can say, for example, there are elements in a lesson plan or home assignment where kids can use AI, and where they can’t.
“If a teacher designs a project the same way she did three years ago, they end up with some kids using AI and some kids resisting it,” she says. This means stopping assignments that can be offloaded onto AI and presented to the teacher as original work.
Classes in the AI era: less tech, more talk
At a practical level, teaching in the AI era may need to be more offline, hands-on and interactive. In-person teaching may need less tech and more talk.
Kallas is a teacher at university level, and she has redesigned assignments for her students, who are training to be teachers.
“For the first outline of a case study, I say to students: discuss with an AI the focus points, such as the topic and what kind of data is needed for this kind of research and prepare with an AI the case study outline.
“Then [the student] comes to class and we have a seminar where we discuss those outlines. They have to defend their outlines in front of class. If they cannot argue, I see they just did five minutes with an AI, and did not focus or learn anything. This is the way you integrate AI into the process, because it forces students much more to think about their outline. AI can give them ideas but they have to regenerate the thought. Then they have to do some empirical research, such as interviewing teachers and students.
“This gets back to arguing in groups, and doing ground research, which cannot be offloaded onto AI.”
At home, this means working with AI as a tutor who forces a student to think, analyse and prepare documents for debate, like a colleague who prepares a student for an exam.
“Memorising, figuring out the new thing, learning about the new thing, understanding the new thing will be done with the help of AI,” she says.
Then in class, rather than rote-learning a new topic, the time is spent analysing what they have learned at home.
Classes then change to include “lots of analysis and lots of teamwork, presentations and discussions, which is mostly developing analytical thinking, critical thinking, systematic thinking, systematising the information and then creating new work.”
Another example is more focus on field work. In a biology class in Estonia, kids go into a park, find flowers and take photos of them, upload the images to ChatGPT as a prompt to give them an analysis. This is arguably a more efficient way of investigating plant life, and kids get to spend more time outdoors, experiencing the subject in its natural environment.
Kallas also uses the example of physics teaching. Formula and rules are hard to understand and memorise in physics, even when taught in class and from books. The teacher can explain them in class, then give a list of prompts to the students, who can use these with the AI at home.
“Kids will be confirming what they have learned at school,” she says, “but the prompt list needs to be given by the teacher. [Education] can go astray if kids use prompts with no guidance.”
Instead, the teacher can say to the kids: ‘We know you are using AI so here is a prompt list’. “Kids have no feeling they are cheating and are learning by using it,” says Kallas.
In general, this style of teaching would require more hours at school because discussion and analysis takes more time than rote learning. “We need more teachers, as we can’t do analysis and discussion with 40 students and one teacher,” she says. “With AI we need more teachers, not fewer.”
In Estonia, she says that teachers do not fear AI will be a replacement, “but they do fear that if they don’t bring in more teachers, learning with AI will become very stressful.”
The digital trap in education: a failure to acquire knowledge
Taking a more assertive approach to technology at this stage of AI development is necessary, due to what happened in the past. In Europe in the 1990s, Kallas says we “got something wrong” when tech was allowed into the classroom.
European schools did not think about the learning processes around that technology, she argues. Since then numeracy, literacy and problem solving skills have been “declining for a generation”, she says.
“When digital technology arrived in the 1990s, Europe [in the education system] started devaluing knowledge learning and memorising. They took knowledge out of the curriculum.”
Learning facts ‘did not make sense’ as they were available at the click of a mouse.
“But learning facts does make sense, such as dates, capitals, languages and history. You need to know when wars started and ended. This was offloaded, and that was a mistake.
“For kids at pre-15, we can’t do critical thinking without knowing the earth is round and not flat.”
Kallas emphasizes that the early years of education still need to focus on knowledge-building.
“Early years from the first grade until grade six should not use AI,” she says.
Risk of burdening teachers with more work
Asking 5,000 teachers to spend more time on reflection on using AI in teaching and redesigning lesson plans could risk becoming an extra burden. This is something Kallas is aware of, and argues it is a question of prioritising. She says school principals have to assign time for learning hubs within the 35-hour working week.
“If this is extra, teachers will not put their energy into this,” she says.
There are a lot of jobs, she says, that teachers should outsource.
“The relationship between home and school in the last 15 years has shifted too much on school and very little on what is done at home,” she says. “When I went to school in the Soviet times, when I didn’t perform well, my parents were invited to school and they were scolded for not doing enough work.
“Nowadays when a kid is not performing, the teacher is invited to the director's office and told; ‘Why are you not doing enough for the kid?’. What about the parents and their responsibility?”
“Teachers have to make sure the kids are mentally well, they have eaten, slept enough, can concentrate and they don’t have discipline problems. [Teachers need] financial literacy, digital literacy, to be psychologists, experts in digital skills and social media, Word, Excel and cybersecurity. We have released parents from their full responsibility. [The system] is too skewed towards the teachers and that’s why they are stressed, because everything is on them.”
Pushback against tech in the classroom
Across the world, many countries have banned mobile phone use in schools, among fears of declining attention spans and cyberbullying. According to UNESCO, 114 education systems (58 percent) now have such a ban in place. Estonia is not among them.
“In this debate about social media and mobile phones, we are talking again about the teachers being responsible for making sure the kids don’t use social media,” says the Minister. “What about the parents? What’s their responsibility?”
Kallas warns against regulating tech “out of the classroom”.
“Regulating phones and digital devices out of the school environment is a wrong direction,” she says. “We can’t teach critical thinking about AI without AI present in the classroom and used for learning. We need to go fast in terms of policy as a learning tool. Education ministers have to be brave among the alarmist voices, otherwise we regulate kids out of the AI generation.”
