This study comprehensively evaluates the current utilization and future potential of ChatGPT in radiology. The primary focus is on its role in enhancing decision-making processes, optimizing workflow efficiency, and fostering interdisciplinary collaboration and teaching within healthcare. A systematic search was conducted in PubMed, EMBASE, and Web of Science databases. Six studies met the inclusion criteria and were included in the analysis, involving a total of 551 chatGPT assessment events. The studies found that ChatGPT output data inaccuracies 80% of the time, with entirely incorrect information in 45% of cases regarding common interventional radiology procedures. ChatGPT performed better on US board-style questions requiring lower-order thinking (P = 0.002) and showed improvements between ChatGPT 3.5 and 4.0 in imaging questions (P = 0.009). The average translational ability score for CT and MRI findings was 4.27/5. While ChatGPT demonstrates substantial potential in augmenting decision-making and optimizing workflow, thorough evaluation and validation are necessary before widespread adoption in radiology. Key limitations include the need for human oversight, ethical considerations, and the absence of strong evidence in image processing. Further research and collaboration are essential to address these challenges and harness the full capabilities of ChatGPT in radiology.This study comprehensively evaluates the current utilization and future potential of ChatGPT in radiology. The primary focus is on its role in enhancing decision-making processes, optimizing workflow efficiency, and fostering interdisciplinary collaboration and teaching within healthcare. A systematic search was conducted in PubMed, EMBASE, and Web of Science databases. Six studies met the inclusion criteria and were included in the analysis, involving a total of 551 chatGPT assessment events. The studies found that ChatGPT output data inaccuracies 80% of the time, with entirely incorrect information in 45% of cases regarding common interventional radiology procedures. ChatGPT performed better on US board-style questions requiring lower-order thinking (P = 0.002) and showed improvements between ChatGPT 3.5 and 4.0 in imaging questions (P = 0.009). The average translational ability score for CT and MRI findings was 4.27/5. While ChatGPT demonstrates substantial potential in augmenting decision-making and optimizing workflow, thorough evaluation and validation are necessary before widespread adoption in radiology. Key limitations include the need for human oversight, ethical considerations, and the absence of strong evidence in image processing. Further research and collaboration are essential to address these challenges and harness the full capabilities of ChatGPT in radiology.