- Гриб'юк, О.О. (orcid.org/0000-0003-3402-0520) (2026) Psychophysiological and didactic aspects of the feasibility and practicality of using information technologies and artificial intelligence in the process of research-based learning in natural sciences and mathematics Габітус (85). pp. 181-190. ISSN 2663-5208
|
Text
HrybiukOO-30.pdf - Published Version Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (221kB) |
Abstract
The study analyses the psychophysiological and didactic aspects of using AI-based information resources in inquiry-based learning; it examines the structure of the appropriateness and practicality of using information resources in the learning process; it analyses models, levels and methods of adaptation in inquiry-based learning in the context of using AI-based information resources for the professional development of teachers, as well as the relevant factors and indicators of influence. The aim of the study is to identify the didactic potential of variable models of the COMSRL and to achieve a thorough understanding of the phenomenon of psychophysiological influence in the context of the appropriate and actual use of information resources and AI in the educational environment during inquiry-based learning of subjects in the natural sciences and mathematics. To achieve the research objectives, the ‘Clever: School of Natural and Mathematical Sciences’ experimental platforms are utilised. A classification of inquiry-based learning methods has been developed, taking into account three variants within the experimental study; accordingly, a comparative analysis has been carried out, considering the specific features of AI use in the inquiry-based learning process. The necessity of selecting AI-based information resources to enhance children’s motivation and level of intellectual development has been established, leading to improved effectiveness in the teaching of natural science and mathematics subjects. The outcomes proved to be statistically significant at a confidence level of p ≤ 0.005. A thorough analysis of the experimental study results indicates that neural networks are susceptible to hallucinations when exposed to rapidly changing content. For a period of several weeks, artificial intelligence models were fed the most mundane content available on the internet, including short videos, Twitter posts, memes, and a variety of news items, amongst others. Consequently, the neural networks exhibited symptoms of 'mental exhaustion' and underwent a decline in performance, evidenced by a 27% decrease in logical reasoning and a 43% decline in the comprehension of long texts.
Downloads
Downloads per month over past year
Actions (login required)
![]() |
View Item |


