Abstract Artificial Intelligence (AI) has become one of the most prominent modern trends in medical imaging, particularly through machine learning and deep learning techniques. This article reviews the general principles of using AI in medical imaging and its applications in image-quality enhancement, lesion detection, classification, workflow support, radiation dose reduction, and computed tomography (CT) image reconstruction. It also discusses major challenges related to data quality, model generalizability, interpretability, ethical considerations, and the need for clinical validation of these systems before their implementation in medical practice. Keywords Artificial Intelligence; Deep Learning; Medical Imaging; Computed Tomography; Image Analysis; Clinical Decision Support. Introduction The availability of large volumes of digital data and the rapid advancement of computational capabilities have significantly expanded the use of artificial intelligence technologies in healthcare. In medical imaging, algorithms can learn patterns and features from medical images and subsequently apply this knowledge to a variety of tasks, including classification, lesion detection, image-quality enhancement, and image reconstruction. These developments have led to the emergence of new applications capable of supporting physicians and healthcare professionals in medical image analysis and improving workflow efficiency while maintaining accuracy, safety, and clinical quality. From Artificial Intelligence to Deep Learning Artificial intelligence encompasses a broad range of methods and technologies designed to enable computer systems to perform tasks that typically require intelligent capabilities. Machine Learning (ML) is a subset of artificial intelligence in which algorithms learn patterns and relationships from data rather than relying entirely on predefined programming rules. Deep Learning (DL), on the other hand, relies on multilayer neural networks capable of learning complex representations from images and data without requiring every feature to be manually designed. This capability has made deep learning one of the most important technologies currently used in medical image analysis. Applications of Artificial Intelligence in Medical Imaging Applications of artificial intelligence in medical imaging include image-quality enhancement and noise reduction, lesion detection, disease classification, image segmentation, quantitative measurements, prioritization of medical cases, and extraction of information that may assist in medical reporting. The benefits of these applications vary depending on the type of data used, the nature of the clinical task, and the quality of the data used to train the model. Therefore, the successful performance of a particular system on a specific task or dataset does not necessarily mean that it will achieve the same level of performance when applied to a different clinical environment or patient population. Artificial Intelligence and Image-Quality Improvement Deep learning algorithms can be trained to reduce image noise, improve overall image appearance, or reconstruct medical images. In Computed Tomography (CT), recent studies indicate that Deep Learning Reconstruction (DLR) techniques can contribute to improving image quality and reducing noise, even when lower radiation dose levels are used. However, the performance of these techniques should be evaluated according to each specific diagnostic task, as an improvement in the overall appearance of an image does not necessarily indicate an improved ability to detect all subtle pathological findings. Artificial Intelligence and Dose Reduction in Computed Tomography Artificial intelligence can contribute to reducing radiation dose in computed tomography through several applications, including improving patient positioning, determining scan range, reconstructing images, and reducing image noise. Recent scientific reviews suggest that these applications may help reduce certain sources of unnecessary radiation exposure. However, the outcomes depend on several factors, including the imaging protocol, type of CT scanner, patient characteristics, and the specific clinical task. Therefore, dose-reduction technologies based on AI should be carefully evaluated before widespread clinical implementation. Scientific Challenges Artificial intelligence applications in medical imaging face several important scientific and technical challenges. These include differences among imaging systems, patient populations, and imaging protocols used across different healthcare institutions. Other challenges include imbalanced datasets, the need for high-quality and accurately annotated data, and the possibility of reduced model performance when applied in environments that differ from those in which the model was originally trained. Some AI-based image reconstruction applications may also produce inaccurate information or artificial structures that do not represent the actual anatomy of tissues, requiring careful evaluation and continuous monitoring. Ethical and Clinical Considerations Medical artificial intelligence systems require appropriate levels of transparency, governance, data protection, and clinical validation. The World Health Organization (WHO) emphasizes that artificial intelligence technologies in healthcare should be developed and used safely and ethically while protecting human safety, autonomy, and fairness in access to and benefit from these technologies. Furthermore, the evaluation of AI systems should not be limited to technical accuracy alone. Their clinical impact, reliability, interpretability, and suitability for the medical environment in which they will be used should also be carefully assessed. The Future Role of the Medical Physicist Medical physicists will play an important role in evaluating artificial intelligence technologies used in medical imaging and verifying their performance and safety. This role includes assessing image quality, measuring noise levels, contrast, and radiation dose, comparing modern algorithms with conventional methods, and verifying the consistency of system performance under different clinical conditions. It will also be necessary to monitor any changes in system performance following software or algorithm updates to ensure the continued quality of medical images, patient safety, and the required diagnostic benefit. Conclusion Artificial intelligence has considerable potential to advance medical imaging by improving image quality, supporting disease detection, enhancing data analysis, and reducing radiation dose in certain applications. However, AI does not eliminate the need for scientific and clinical evaluation, nor does it replace specialized human expertise. The integration of artificial intelligence into medical imaging should be based on independent performance validation, high-quality data, continuous system monitoring, and a clear understanding of the capabilities and limitations of these technologies. Through collaboration among physicians, medical physicists, artificial intelligence specialists, and biomedical engineers, these technologies can be utilized more safely and effectively, contributing to improved healthcare quality and greater accuracy and efficiency in medical imaging. Almustaqbal University – The First University in Iraq