Generative Artificial Intelligence (genAI / GAI) is gaining ground in various industries, and education is no exception. The use of AI in education can not only increase the efficiency of teaching, but also open up new possibilities for personalized learning. In this post, we explore how (generative) AI can be used in education, its benefits and challenges, and prospects. The timeliness of the topic is given by OpenAI’s announcement of the availability of ChatGPT edu, a version of the chatbot designed for educational environments.

The initial application of AI technologies in education was mainly focused on the automation of administrative tasks and the development of e-learning systems. These were basic tasks that could be managed relatively reliably by solutions of the time. These systems helped, for example, to manage enrolment processes, monitor student performance, and produce timetables, thus saving considerable time and money. Current development trends are much more far-reaching. Just think of the announcement by Udemy, who promise to use AI to offer adaptive learning pathways. They are tailored to the individual needs of students, thus increasing learning efficiency. The two main selling points are usually the adaptive learning system itself and the possibility to provide automatic, personalized feedback.

The former suggests that AI-based e-learning platforms can analyze students’ learning patterns and recommend personalized learning paths. For example, if a student is struggling with a particular topic, AI can offer more practical exercises and explanations. The essence of the latter is that such systems can provide immediate, personal feedback to students after completing tests and assignments. This allows students to immediately learn where they made mistakes and how to improve their performance. This can, for example, prevent incorrect patterns from becoming locked in during the learning process, shortening the time needed for subsequent correction. Without being exhaustive, let’s look at some of the pros and cons of this topic.

Arguments for the use of AI in education…

As mentioned above, proponents say that one of the biggest benefits of (generative) AI in education is the potential for personalized learning. AI can analyze students’ learning styles, strengths and weaknesses, and then create tailored learning material based on this. This can be particularly useful for students with special needs.

Assessment has been present in education since the very beginning, but technological advances have transformed the role of the concept. Rather than simply measuring students’ memorization skills, assessment is becoming more of a catch-all term. It also involves the use of data to make strategic decisions, improve curricula, and optimize learning experiences.

Artificial intelligence has become an important tool for providing immediate feedback and practical opportunities for both students and teachers. Online platforms such as Gemini, ChatGPT or even Claude can generate simulated evaluation scenarios and provide instant feedback. This can help students prepare for real assessments and encourage self-evaluation. The other side of the coin is that educators can use AI to refine their assessment methods, ensuring that their questions are effective and accurately measure learning outcomes.

In essence, we can say that AI acts as a kind of practice space, creating a feedback loop that improves the quality of assessments and aligns teaching methods with learners’ understanding of the text. This of course includes all the activities that are usually referred to as traditional curriculum development.

These are mainly in terms of the learning experience, but of course, there is also a significant economic dimension to the use of AI in education. In dry terms, the introduction of AI technologies in education can lead to cost savings by automating administrative tasks and optimizing teaching processes. Ideally, the resources freed up can be used for activities such as monitoring students’ mental health and improving the educational “experience”.

… and against

Despite the billions invested in AI technologies, the problem of lack of trust in AI in general continues to pose unending challenges. The general (often not unfounded) fears that surround the introduction of such technologies in most areas also arise in the field of educational applications.

The lack of trust in AI is caused by a combination of (real and perceived) risks. These include disinformation, security problems, the black box problem, ethical concerns, biases, instability, hallucinations of Large Language Models (LLMs), unknown risks, job losses, environmental impacts, industrial concentration, and state overreach. The latter is a particularly significant risk for authoritarian regimes. These risks combine to create widespread skepticism and business concerns that hinder the adoption of AI. For example, radiologists are reluctant to use AI when the opacity makes it difficult to understand how the algorithm makes decisions during medical image analysis.

There are three main implications of the above concerns. First, despite improvements in AI performance, AI users still face an insurmountable lack of trust. Second, companies need to invest in understanding and managing the biggest risks. Third, the pairing of AI and human skills will be the most important risk management tool, which means that humans will always be needed to help bridge this gap, and humans must be properly trained for this task.

The latter is a particularly important aspect of education, since, especially in primary and secondary schools, the transfer of knowledge is not the only task. Social integration, learning the rules of social coexistence, and the development of emotional intelligence are also prerequisites for the development of quality human relations. One of the most important tasks of teachers is to prepare students to deal with challenges, to manage stress and to set important goals in life.

This is precisely why human interaction is essential in education, because teachers not only provide information, but also empathy, support and motivation for students. Teachers can sense the emotional and mental state/well-being of students and provide personalized help, something that AI cannot currently do. In addition, the development of critical thinking and creative problem-solving requires human guidance.

In addition, a major challenge for the deployment of AI technologies is to control the use of data. Important decisions will also need to be made at all levels of society about the ownership of data and the best ways to use it in a transparent and ethical way. Of course, it may seem excessive to talk about an Orwellian future, but we cannot forget that the prevention of such dystopias is the responsibility of all of us, not tomorrow, but today!

Although AI offers exciting developments, particularly in improving education worldwide, we are still in the early stages of its use. It is certainly true that many experiments and research are needed before widespread application to successfully introduce AI tools in higher education institutions. In addition, of course, students need to learn how algorithms use data for decision-making and how they themselves can use the results of these algorithms in a responsible way. Above all, it must be ensured that students remain informed about how and for what purposes their data is used so that they can intervene if necessary.

István ÜVEGES is a researcher in Computer Linguistics at MONTANA Knowledge Management Ltd. and a researcher at the HUN-REN Centre for Social Sciences, Political and Legal Text Mining and Artificial Intelligence Laboratory (poltextLAB). His main interests include practical applications of Automation, Artificial Intelligence (Machine Learning), Legal Language (legalese) studies and the Plain Language Movement.