The impact of artificial intelligence technologies on increasing the efficiency of distance education in technical universities during global challenges

Abstract

The study substantiates the role of integrated artificial intelligence technologies in improving the effectiveness of distance education at technical universities in the face of global challenges and a shortage of laboratory practice, formulating the goal to quantify the impact of adaptive platforms, intelligent assistants, automated verification, proctoring and predictive analytics on academic performance, student engagement and teacher time allocation. In 2020-2022, 3,450 students were examined at three technical universities (control 1,720; experiment 1730) and 410 teachers; the experimental group was trained using Al tools, and the control group was trained in a standard LMS; exam scores in basic disciplines, logs of activity in the LMS, the distribution of teachers' labor costs and satisfaction surveys were collected, t criterion, U criterion, correlation analysis and ANOVA were applied. A significant increase in academic results was recorded: higher mathematics 3.844.38, general physics 3.91 4.29, fundamentals of programming 4.02 4.55 at $p 0.001$ and a decrease in the spread of grades, which reflects the alignment of the level of training and targeted support for the risk group using predictive analytics. Engagement has increased significantly: average LMS time is 6,458,92 hours per week (+38,29%), interaction with content is +47,43%, the share of active forums 18,33% 35,19% (+91,98%); A relationship was found between forum activity and the final score $(r=0.72)$. There was a radical optimization of the teachers' work: task verification 35,82% 14,25%, answers to standard questions 15.47%→3.16%, with an increase in the share of material development 29.15%→48.93% and individual consultations 19.56%→33.66%. Student satisfaction increased in all parameters, including timely feedback (+1.74 points) and personalization (+1.88) on a 5-point scale; the combined Al factor explains over 65% of the variation in the final indicators. The conclusion is made about the systemic and synergetic effect of the Al complex, which makes it possible to compensate for the methodological limitations of distance learning in engineering education; prospects are associated with the introduction of Al-controlled virtual laboratories, the expansion of predictive models and compliance with ethical and legal standards of working with data.