The specifics of the effectiveness of reading complex professional texts when working with an AI agent: Assessment of the level of assimilation, taking into account the establishment of the duration and amount of gaze fixation
Authors:
- Lyubov Nikolaevna GladkovaLGLyubov Nikolaevna Gladkova
- Olga Valerievna KrezhevskikhOKOlga Valerievna Krezhevskikh
- Dmitriy Viktorovich SkorobogatoDSDmitriy Viktorovich Skorobogato
- Anastasia Alekseevna SavelyevaASAnastasia Alekseevna Savelyeva
- Olga Vladimirovna BulatovaOBOlga Vladimirovna Bulatova
- Anastasia Ilyinichna GafnerAGAnastasia Ilyinichna Gafner
Open Access
- 28 February 2026
- journal article
- Published by Novosibirsk State Pedagogical University in Science for Education Today
- Vol. 16 (1) , 246-270
Abstract
Introduction. The article raises the question of the effectiveness of reading complex professional texts in the field of education using a traditional approach and working with an AI agent. The purpose of this article is to identify the level of material acquisition by undergraduate students when reading complex professional texts in the field of education and working with an AI agent, taking into account the duration and number of fixations on the material. Materials and Methods. The research is based on the theory of the activity–based approach (A.N. Leontiev), in which cognition is organized on the basis of intensive cognitive activities of students in the process of formulating and researching their prompt requests to the AI agent. The design of the study assumed measurements of indicators in the experimental and control groups. Testing was used to establish the level of memorization of large texts and the level of positive motivation when reading complex professional texts in experimental and control groups. The technologies of attention fixation (‘eye tracker’) were used, as well as video recording of the student’s work with the text. The level of readability of the text, its linguistic simplicity and clarity of presentation, the length of sentences and the complexity of syntax, as well as the structure of paragraphs and the division of the text were assessed in advance through AI systems of Text Meters on a level scale. At the end of the experiment, the results of filling in knowledge maps by participants of the control and experimental groups were analyzed, as well as the reports for fixing students’ gaze on the text under study were evaluated. The Pearson correlation coefficient was used to establish the presence of a statistically significant relationship between variables such as gaze fixation on the educational text and the final educational results of students in mastering the topics studied. Results. It was revealed that the duration of fixing students’ gaze on the text of the original source is less than on the text generated by the chat bot; the number of students fixing their gaze on the material when working with an AI agent exceeds the same indicator in the group working with a traditional text source by 45.3%; there is a statistically significant direct relationship between the fixation of the gaze on the text and the level of assimilation of the studied material (the value of the Pearson correlation coefficient was 0.7, which indicates the presence of a moderate positive correlation between these parameters); when working with an AI agent, a high (meaningful) level of assimilation of educational material prevails, recorded in the majority of subjects (54.55%). On the contrary, the group working with the traditional primary text source is dominated by a predominantly low (fragmentary) level of learning of the educational material. Conclusions. The authors concluded that learning material is more effective when working with an AI agent than when reading complex professional texts in the field of education, but emphasize the importance of further research to compare the results of material acquisition when working with an AI agent and other innovative technologies, including those based on AI.Keywords
This publication has 25 references indexed in Scilit:
- AI-driven language learning in higher education: an empirical study on self-reflection, creativity, anxiety, and emotional resilience in EFL learnersHumanities and Social Sciences Communications, 2025
- Qualitative research of students’ cognitive preferences when using artificial intelligenceScience for Education Today, 2025
- Chatbots and student motivation: a scoping reviewInternational Journal of Educational Technology in Higher Education, 2025
- Students’ and teachers’ attitudes towards the use of tools with generative artificial intelligence at the universityThe Education and science journal, 2024
- Risks and Challenges in Introducing Artificial Intelligence into Higher Education2024
- Using Chatbots to teach English as a foreign language: A systematic literature review from 2010 to 20232024
- Understanding self-directed learning in AI-Assisted writing: A mixed methods study of postsecondary learnersComputers and Education: Artificial Intelligence, 2024
- Is it harmful or helpful? Examining the causes and consequences of generative AI usage among university studentsInternational Journal of Educational Technology in Higher Education, 2024
- A comprehensive AI policy education framework for university teaching and learningInternational Journal of Educational Technology in Higher Education, 2023
- Integration of artificial intelligence performance prediction and learning analytics to improve student learning in online engineering courseInternational Journal of Educational Technology in Higher Education, 2023