The most common method of diagnosing knee injuries is magnetic resonance imaging (MRI); however, interpreting MRI results can be inconsistent and time-consuming. Finding general abnormalities and specific diagnoses such as meniscal tears and Anterior Cruciate Ligament (ACL) tears was the main goal. Maximizing efficiency and reducing diagnostic error rates was the ultimate aim of automating the process of identifying abnormalities and specific knee injuries. The ability to detect ACL tears and meniscal tears is crucial, as these are common knee injuries with significant clinical implications. The study likely involved training the deep learning model on a diverse dataset of knee MRI exams to ensure robust performance. Integrating the model into clinical workflows could streamline the interpretation process and facilitate prompt intervention for patients with identified abnormalities. Providing clinicians with the model’s predictions during interpretation could lead to more informed decision-making and improved patient outcomes. This approach may contribute to a more efficient healthcare system by optimizing resource allocation and reducing unnecessary delays in diagnosis and treatment. The model achieved accuracy of 0.862 for abnormal, 0.87 for ACL tear, and 0.79 for meniscus tear.