earticle

영어

Recent advances in artificial intelligence (AI) have introduced a new turning point in the field of rehabilitation training. In particular, AI robot-based training has attracted attention by enabling patient-centered rehabilitation through precise biomechanical data analysis, motion correction, personalized exercise programming, and real-time feedback. In this study, a rehabilitation training program was designed based on the structure and functionality of the AI training equipment Moty, through collaboration among a team of experts consisting of AI developers, fitness trainers, sports education researcher, and rehabilitation professionals. The effectiveness of the program was qualitatively analyzed using the four levels of the Kirkpatrick model—reaction, learning, behavior, and results. Data were collected through participant interviews, rehabilitation journals, video analysis of training sessions, and expert consultations, and inductive category analysis was conducted using the constant comparative method. The results revealed that, first, in the reaction phase, Moty's automated weight adjustment and real-time feedback promoted psychological stability and engagement, fostering learners' sense of novelty and immersion. Second, in the learning phase, participants reported embodied understanding of muscle-specific stimulation through concentric, eccentric, and isokinetic modes, leading to deeper comprehension of exercise principles. Third, in the behavior phase, Moty's simplified setup and automatic data logging reduced physical burden and injury risks, encouraging more autonomous and consistent exercise behavior. Fourth, in the results phase, improvements in functional recovery and sustained rehabilitation participation were observed. Notably, the emotional support and continuous attention provided by rehabilitation professionals compensated for the limitations of AI technology, reinforcing participants' trust and motivation. Meanwhile, some limitations of technology-centered instruction were noted, including operational errors of the AI robot, difficulties in interpreting feedback, and mechanical constraints. In addition, the effectiveness of AI robot-assisted rehabilitation appears to be significantly enhanced when accompanied by human-centered interaction and empathetic communication.