The results indicate that radiography education is beginning to adopt AI tools in radiology. Both graduates and students demonstrated a satisfactory level of knowledge about AI applications in medical imaging, particularly in region segmentation. Furthermore, the findings suggest a strong interest among students and graduates to further develop their AI knowledge and skills within radiology. This highlights the potential for AI to enhance patient care and outcomes in medical imaging, while also meeting the evolving demands placed on radiography professionals [29]. Integrating AI simulation tools into the learning process could therefore support students in achieving better educational outcomes in radiography training.

Given that students and graduates had not received formal education or structured information about AI, many likely had limited or no understanding of its applications. As a result, the survey design incorporated only Yes/No response options to reduce uncertainty and enhance clarity in the data collected. While this approach has limitations, it allowed for more consistent interpretation of responses. Participants regarded the enhancement of AI simulation technology in medical imaging as a valuable contribution to radiography education [30]. The post-lecture questionnaire revealed a shared view among students and graduates that AI simulation offers significant benefits for learning. However, graduates expressed concerns regarding the current lack of realistic clinical images within AI simulation tools used for education. The results emphasized that the case images used in experiments were insufficiently advanced to replicate real-world clinical scenarios accurately [31]. This disparity likely stems from graduates’ greater clinical experience compared to students and underscores the need for more sophisticated case simulations.

Graduates recommended improving and expanding the current AI simulator, which is primarily used for teaching basic human anatomy in radiography training. Following the lecture, they acknowledged that generative AI holds promising potential to optimize the teaching process and enhance radiography education. Nevertheless, graduates also highlighted the importance of addressing issues related to the stability and reliability of generated images [31]. As a result, they expressed a desire to actively participate in the research and development of generative AI applications in radiography education, as reflected in the findings. These findings suggest that AI-based educational tools should be integrated gradually, with attention to aligning simulation fidelity with the clinical experience level of learners [31]. For early-year students, simplified case studies focusing on anatomical recognition and basic protocols may be most appropriate, while more advanced simulations could be introduced in later stages to reflect real-world complexity.

Integrating ChatGPT into online learning systems offers a more interactive and automated approach to radiography education, potentially enhancing the precision and effectiveness of educational outcomes [5, 24]. The in-house AI imaging simulator can generate detailed medical images based on extensive textual input; however, students with limited clinical experience may find this challenging. This AI imaging simulator could provide tailored and effective learning interactions, enabling students to acquire foundational knowledge of anatomy and medical imaging for rare diseases. The radiography education program aims to deliver proper training for healthcare professionals, particularly those in regional areas who face barriers to accessing educational institutions due to time and resource constraints [32]. Online learning could offer a viable solution for medical imaging professionals in these regions to acquire essential knowledge and skills [20]. This flexible approach allows students to participate remotely from home or the workplace, making it well-suited to their needs [20].

Moreover, graduates often seek more profound insights into the applications of AI in medical imaging technology. While the requirements of graduates and students may differ, their goals align in expanding their understanding of AI’s potential and future applications in clinical practice [33]. Such preparation is crucial, given the demand for timely decision-making and continual learning in evolving medical environments [34]. Additionally, online learning platforms facilitate knowledge sharing between lecturers and students, supporting ongoing professional development for medical imaging practitioners in regional or remote settings [19, 21]. This enables them to stay updated with the latest technologies and skills.

Finally, the findings underscore the importance of intensifying efforts to develop the in-house AI imaging simulator. Encouragingly, students and graduates demonstrate a willingness to engage in this developmental process. Involving students in the creation and refinement of such simulators may prove highly beneficial in enhancing their comprehension and practical experience within radiography education [31].

Since the release of ChatGPT, the capabilities of generative AI technology have advanced significantly. However, there is a tendency to overemphasize the current potential of AI in imaging simulation, particularly in medical imaging applications [24]. Graduates now have greater access to AI tools in the workplace, which raises their expectations for professional practice. AI simulator tools can help graduates navigate the evolving challenges and opportunities within clinical medical imaging [30].

AI technology used in medical imaging can be extended to clinical disease simulation, thereby enhancing both practical skills and theoretical knowledge [17, 29]. In contrast, students still pursuing radiography education often have more modest expectations due to their limited clinical exposure. Since students may not yet have developed practical skills or applied their knowledge in real-world scenarios, the transition from academic study to clinical practice can be daunting. Consequently, the demand for AI simulation tools may differ between students and graduates. AI simulators provide students with the opportunity for hands-on experience and exposure to realistic clinical cases, potentially bridging the gap between their expectations and reality [29, 31].

The results suggest that both students and graduates lack full confidence in the transformative impact of AI on medical imaging. This skepticism may reflect doubts about the promises made by AI proponents, as many current implementations rely on limited datasets that constrain the effectiveness of AI models. Despite the widespread adoption of machine learning and cognitive computing across various applications, concerns persist about the efficiency and reliability of AI tools. Although AI has demonstrated capabilities in performing highly technical tasks, its dependability is often questioned, especially when precise and objective outputs are required. Factors such as data quality, algorithm design, and unpredictable variables contribute to this unreliability. The significant gap between expectations of AI’s potential and the limitations posed by constrained data and models likely explains the skepticism observed among students and graduates.

A foundational understanding of human anatomy remains essential in medical imaging education [35, 36]. While there is room for improvement in AI’s ability to simulate complex clinical cases, AI simulators can significantly increase student engagement by providing opportunities to practice clinical imaging across multiple modalities [29, 31]. AI simulators could have a significant potential to reduce resource barriers in rural areas, facilitating fundamental education and practical experience for radiography students. Survey and questionnaire findings indicate a growing interest in the application of AI within continuing professional development (CPD) education [37]. This technology could transform the teaching of human anatomy, making it more effective and accessible, especially for learning about rare diseases [31, 35].

Clinical databases and medical images are often inaccessible to the public, limiting their use for training machine learning models. Consequently, AI-generated images may not accurately resemble real clinical images. This lack of realism can increase the risk of producing unrealistic simulations, thereby reducing the effectiveness and relevance of AI models in both clinical and academic settings. To enhance the accuracy and reliability of AI tools in medical imaging education, access to comprehensive and diverse clinical databases is crucial.

Moreover, recent studies suggested that AI simulators could generate accurate and feasible clinical scenarios using only textual inputs [5, 24, 38]. This text-based simulation approach holds promise for future development, offering a cost-effective and efficient alternative to image-based simulations. However, more complex deep learning models and expansive clinical datasets are necessary to further advance image simulation technologies. Continued improvements in AI imaging simulation will enable ongoing exploration and innovation in this area.

For AI-driven tools to effectively address realistic medical challenges, a unified and comprehensive body of knowledge is required to cover a wide range of clinical scenarios [33, 35]. Although the proposed online learning platform is designed to function in both online and offline modes, larger offline clinical image databases are needed to enhance image generation capabilities.

One major challenge in integrating AI simulators into online learning platforms is ensuring the protection of patient information. Collaborating with open-source clinical datasets may offer a viable solution, while also addressing legal issues related to copyright infringement [39]. When incorporating APIs with large language models (LLMs), it is essential to provide clear information regarding image copyrights. Only images with proper copyright clearance should be used, or permission must be obtained from copyright holders.

While conventional clinical image simulations remain useful for teaching, AI-driven simulation tools offer greater convenience and effectiveness. Such tools can enhance student engagement and participation, providing a robust platform for knowledge acquisition and retention. This, in turn, better prepares students for advanced studies and clinical practice. Combining AI simulators with self-learning LLM models, such as ChatGPT, could create a transformative learning environment for radiography students, ultimately enhancing their practical performance [5, 24]. The ongoing development of AI technology presents a promising new approach to radiography education. Integrating AI simulators and self-learning LLM models into online learning systems could be a new approach to training radiographers, potentially enabling students to learn more effectively and thoroughly.