AI for swimming recommendation systems exploring the current landscape and research opportunities

Artificial intelligence (AI) is increasingly being integrated into sports analytics, improving technique analysis, performance prediction, and training personalization. However, in aquatic disciplines such as swimming, AI adoption remains limited. This paper systematically reviews 42 peer-reviewed studies (2018–2024) at the intersection of AI and swimming to map current progress and research gaps. The analysis reveals that while over 80% of studies focused on stroke classification, turn detection, or fatigue monitoring, none implemented full recommendation systems for swimmers. Performance gains reported across these studies range from 90 to 99% accuracy in stroke recognition tasks using CNN-LSTM and BiLSTM architectures. To address the lack of integrated solutions, this paper outlines a three-layer framework that combines multimodal sensing, contextual awareness, and adaptive feedback for real-time, swimmer-centric recommendations. By consolidating the state of research and proposing a structured development roadmap, this work aims to guide the design of intelligent, evidence-driven recommendation systems for the swimming domain.