Basketball's global popularity is evident in its international presence and commercial collaborations. Shooting, a fundamental offensive skill, is crucial for scoring and winning. Artificial Intelligence (AI) integrating into players’ training, such as tracking, analysis, and instruction, is rapidly expanding. The National Basketball Association (NBA) utilizes AI through multi-angle video analysis, game statistics, and wearable technology to enhance training and performance. This demonstrates the potential of AI in professional sports. However, general college basketball teams often lack access to such technology, particularly in basic skill development like shooting. This study employed an affordable AI-powered app to assist shooting training for college basketball players. This app could provide real-time feedback and video analysis, allowing players to correct their postures and improve accuracy via using their own cellphones. It also could offer valuable training insights for coaches.

The purpose of this study is to investigate the effectiveness of the AI-assisted shooting training app, HomeCourt, for college basketball players. There were six basketball players from a university in northern Taiwan, aged 19-23, as the participants in this case study. They were divided into a high-experience group (over 3 years of training) and a low-experience group (less than 1 year of training). This study took two months with six weekly assessments. Each session followed a standardized procedure: 10-minute warm-up, 10-minute pre-test, 15-minute video review and discussion, 13-minute shooting practice, 2-minute hydration break, and 10-minute post-test. The app was used for all shooting practices to make real-time recording and track accuracy, and it could generate performance charts and store videos for detailed analysis. Players could review their postures, identify aspects for improvement, and check their shooting accuracy from various court positions.

Following the sessions, the Situational Interest Scale was administered, and interviews were conducted to gather further insights. The results showed that the accuracy of both groups had been improved, especially for the high-experience group. The pre-test averages for the high-experience group were 50%, 54%, and 62%, while the post-test averages were 52%, 58%, and 73%. Meanwhile, the pre-test averages for the low-experience group ranged from 32%, 34%, and 39%, with post-test averages of 35%, 37%, and 39%. Results of Situational Interest Scale showed that the attentional demand had the highest with a mean score of 4.63. The lowest-scoring dimension was challenge, with a mean score of 1.46. The attentional demand dimension suggests that AI-assisted shooting training could enhance concentration during practice.

Participants generally agreed that this training method was unique and stimulated their practice motivation during the interviews. The results were consistent with the previous literature. In conclusion, AI-assisted shooting training effectively improved players' shooting accuracy with positive feedback, indicating that AI holds promise as a valuable tool for future basketball training or other physical education. Discussion and Suggestions were also provided.

Keywords: Artificial Intelligence, Physical education, Basketball training, Shooting training.