The human movement assessment is a crucial element of physical rehabilitation centers. It enables the assessment of exercise quality, patient progress, and therapy efficacy. Traditional rehabilitation assessment mostly relies on therapists’ manual observation, which can be subjective, time-consuming, and challenging to apply consistently across diverse professional and home settings. Therefore, artificial intelligence (AI) has enabled the development of automated systems that objectively evaluate movement. These methodologies are systematically classified into sensor-based techniques utilizing inertial measurement units (IMUs) and vision-based approaches dependent on computer vision and posture estimation. Despite the promising results from these systems, many unresolved challenges remain. The present methods exhibit insufficient robustness in real-world rehabilitation environments affected by occlusions, variations in illumination, and ambient clutter. Secondly, IMUs are precise yet uncomfortable for patients. Moreover, there is an absence of systems that identify hand movements due to challenging recognition. This research offers a comparative analysis of AI-driven tools for assessing movement in rehabilitation. It defines the gaps and the essential future research opportunities.