For partners 
Artificial intelligence in education has outgrown the experimental stage. Following the widespread adoption of generative neural networks, schools, colleges, and universities faced a practical question: how to use a technology that is already capable of writing essays and compositions (and suggesting the difference between them), solving problems, explaining educational material quite well, checking texts, creating tests, and taking on part of a teacher's routine work?
Different countries' approaches vary considerably. Some states focus primarily on building a regulatory framework and seek to ensure oversight of AI use, others bet on the rapid introduction of technology into schools and universities, while others include developing AI literacy among the priorities of national educational policy.
One of the key topics of the article is the adaptation of education to the AI reality. The discussion is no longer only about whether a schoolchild may use ChatGPT when doing homework; today the focus has shifted to other questions: what knowledge and skills children need to be taught in the age of neural networks, how to prepare teachers to work with new technologies, who bears responsibility for algorithmic errors, how to ensure the protection of students' personal data, and where the line lies between using AI as a support tool and losing independent thinking and freedom of will.
The gap between supply and demand
The need for AI knowledge and skills is growing faster than educational systems can adapt. According to the Stanford AI Index 2025, 81% of computer science teachers believe that introducing students to artificial intelligence technologies should already be part of the basic computer science curriculum. However, educators themselves are not always ready to teach such topics: only 46% of high school teachers feel at home with AI, 44% of middle school teachers, and 34% of elementary school teachers.
This clearly illustrates the main contradiction of modern educational policy: students are already actively using neural networks, employers expect graduates to have AI skills, universities are rethinking their approaches to assessment, while teachers are living in the realities of the past decade, if not century.
There is also a global inequality: while some countries are already discussing the possibility of giving every student a personal AI assistant, in others many schools still lack stable internet access, computers, and even electricity. Therefore, it is premature to speak of a single global scenario for introducing AI into education. Different countries are approaching these changes in different ways.
The European model: regulation first, then implementation
The European approach to AI in education can be called the most cautious. The EU seeks to develop its own technologies and strengthen its position among world leaders in AI, while simultaneously building a detailed legal regulatory framework for this purpose.
The key document here is the EU Artificial Intelligence Act, adopted in 2024. It classifies some AI systems in education as high-risk; this applies, for example, to technologies that may influence a person's access to education, assessment of results, allocation to educational pathways, or important decisions about a student. Such systems are subject to strict requirements regarding data quality, risk assessment, transparency of operation, documentation, and ensuring human oversight: if an algorithm can influence a student's educational trajectory or future, it cannot be treated as an ordinary digital tool.
A separate direction is the formation of AI literacy. The EU Act obliges organizations using artificial intelligence systems to ensure the necessary level of preparation for employees and other users of such technologies. This means that a teacher, administrator, methodologist, or university lecturer can no longer simply learn to use a neural network at a basic level — it is important to understand its capabilities and limitations, know what data it operates on, in which cases it may make errors, and what risks are associated with its use.
At the same time, the European approach is not limited to regulation and restrictions. In April 2025, the European Commission presented the AI Continent action plan, aimed at developing artificial intelligence infrastructure, preparation of personnel, and broader implementation of AI technologies in the economy. For the education sector, one of the key elements of this strategy was the AI Skills Academy — an initiative designed to support university programs, preparing and retraining of specialists, and to develop mentoring and internship systems in companies.
At the school education level, a unified approach in Europe has not yet taken shape: each country is developing its own strategy. One of the most notable examples is Estonia and its AI Leap program: in the first phase of the project, around 20,000 high school students and 3,000 teachers gained access to AI-based tools, and the program was subsequently planned to be extended to other groups of students and educators. For a small country, this is essentially an attempt to integrate AI into everyday school practice in the same way that computers and the internet were integrated before.
But in general, the European approach remains cautious; AI here is viewed as a useful learning tool whose potential requires clear rules and oversight. At the same time, it is considered that the risks of errors, discrimination, data breaches, and non-transparent algorithmic decisions are too great for such technologies to be deployed without proper supervision.
The American model: market, speed, and technological leadership
The United States is moving in a different direction. The American model is far more closely tied to the market, technology companies, and the logic of global competition, so AI is viewed not only as an educational tool, but also as a factor of economic and geopolitical leadership.
In January 2025, the Trump administration issued an executive order removing barriers to American leadership in AI. In April, a separate executive order on AI education for youth appeared: its goal is to develop AI literacy, prepare teachers, introduce students to technologies earlier, and build the future workforce for an economy where AI will be a basic tool.
At the federal level, the focus is on general guidelines rather than a detailed program for schools. Practical decisions are made at the level of states and school districts, universities, unions, foundations, and technology companies. This reflects the approach characteristic of the United States: the government sets the strategic direction, while implementation and expansion of initiatives are largely ensured through partnerships.
The most illustrative example is the National AI Teaching Academy, being launched by the American Federation of Teachers, OpenAI, Microsoft, Anthropic, and the United Federation of Teachers. The stated goal is to prepare 400,000 school educators over five years — that is, approximately one in ten K–12 teachers in the United States. The program involves in-person and online courses, hands-on sessions, work with real lesson scenarios, and support for educators after studying.
The point of this initiative is not only to teach teachers how to use a chatbot. Far more important is something else: returning the educator to the center of decision-making. American education has already seen cases where new technologies were introduced without the real participation of teachers, and with AI, unions and part of the educational community are trying not to repeat that scenario: the teacher must understand the tool, see its limitations, and participate in choosing the rules.
At the same time, the American approach has obvious risks. When large technology companies begin to play a key role in developing educational tools, methodologies, and studying programs, schools gain faster access to new solutions, but their dependence on market players also grows. In such a situation, questions inevitably arise about the protection of student data, the influence of commercial interests on the educational process, algorithmic transparency, and who ultimately defines the rules of the game in education — the teaching community or BigTech.
The American model is built on speed: the prevailing view is that overly strict regulation can slow development, causing America to lose its advantage. Therefore, the emphasis is not on restricting AI, but on its active integration into the educational system. Possible risks and regulatory measures are discussed, but they concern civil rights advocates more than the professional community and policymakers.
The Chinese model: everything goes according to plan
The Chinese strategy differs from the European and American ones primarily in its degree of centralization: AI here is viewed as part of a large project of technological sovereignty, transforming the country into a leading educational and scientific power.
In 2025, China announced its intention to embed artificial intelligence into all levels of education — from school curricula to university disciplines, textbooks, and teaching methods. This initiative fits into a broader national strategy: by 2035, China aims to become one of the world's leading educational powers and to prepare the specialists needed for the development of a high-technology economy.
The Chinese approach is centralized: the Ministry of Education sets the framework, and schools and universities are required to follow it. The option of «waiting for teachers to figure it out on their own» is practically not applied here. AI must become part of the educational process from the earliest stages. Children are taught to understand how intelligent systems work, where they are applied, what tasks they solve, and what their limitations are; scenarios for using AI in various educational situations are developed for educators: when explaining material, preparing assignments, organizing project work, and analyzing educational data.
A separate direction is the use of AI in school management: technologies are used to automate routine tasks, plan events, analyze data, identify risks, work with archives, and support school life.
At the university level, China is expanding and updating programs in areas such as the digital economy, engineering, medicine, agriculture, the humanities, and high-technology industries. At the same time, the task is not limited to preparing programming specialists: AI technologies are gradually being integrated into a wide variety of fields — from medical diagnostics and biotechnology to the agro-industrial sector and urban management.
The strength of the Chinese model is speed and systemic consistency. If the government decides to introduce AI into education, it can quickly restructure curricula, teacher preparation, administrative requirements, and university programs. The weakness is the risk of excessive centralization and control: the deeper AI enters school management and the analysis of student data, the more sensitive (and acutely felt) the issues of privacy, transparency, and the boundaries of surveillance will become.
The main challenge is not AI itself, but how learning is changing
AI affects education not simply as another digital tool — its emergence forces a rethinking of the very organization of the learning process. If a neural network is capable of writing an essay or explaining a new topic, the educational system is forced to find new answers to fundamental questions: what counts as knowledge under new conditions, how to assess a student's independent work, what to focus on during lessons, what assignments to give for independent completion, how to evaluate learning outcomes, and which competencies will remain the prerogative of humans.
For educators, this means the emergence of new professional tasks: it is important, necessary to understand the capabilities and limitations of neural networks, to recognize their typical errors, to explain to students the principles and rules of working with such tools, to ensure data protection, to adapt learning assignments — and at the same time not to reduce the lesson to a demonstration of the latest technological novelty.
AI literacy in this regard is becoming a new fundamental competency — roughly the same as digital literacy was 20 years ago. At first it was perceived as a set of specialized skills for a narrow circle of professionals, but gradually it became necessary for almost everyone. The only difference is that today's changes are happening much faster, and there is almost no time for gradual adaptation.
