Purpose:
Visual tools, including graphic narratives and storytelling, are increasingly used in medical education and patient care to enhance understanding (Maatman et al, 2019). In fields as nutrition and mental health, these visuals have proven effective (Branscum et al, 2013). With the rise of Artificial Intelligence (AI) image generators, new opportunities are emerging for creating content, though concerns around functionality, authorship, and copyright persist (Hickey, 2023). While court rulings are pending, the U.S. Copyright Office requires human authorship for copyright protection, and the degree of human input in AI-generated content remains a key consideration (MacDermott, 2023). AI-generated images can create representations of patient scenarios without needing real patient photographs, thus avoiding HIPAA (Health Insurance Portability and Accountability Act) violations. This study evaluates AI-generated images for patient education on headache, aiming to identify which systems produce the most useful visuals for improving learning in healthcare.

Methods:
Three publicly accessible AI platforms Gemini, Whisk, and OpenArt were used to generate images based on identical prompts of primary headache disorders. The initial prompt was: “Create an image of a 24-year-old female suffering a migraine headache episode.” A second, more complex prompt “Create an image of a 62-year-old male with a cluster headache” served to assess each system's response to limited details. The third prompt was to “Create an image of a 35-year-old office worker with a tension-type headache characterized by bilateral tightness”. When options were available, the most realistic style was selected. The study is part of a broader investigation into digital tools for pain management education and is IRB-exempt (UP-24-00146).

Results:
All three systems were user-friendly, with intuitive interfaces for inputting prompts. Gemini generated a single image; Whisk provided two images and allowed optional animation; OpenArt offered a broader selection, generating up to eight images and showing examples of possible outputs. All platforms produced visuals relevant to the prompts, though differences in style and interactivity emerged. Whisk stood out for its animation options and customizability. When tested with the second prompt, both Gemini and Whisk produced cartoon-style images, similar in appearance and aligned with the request.

Conclusions:
AI-generated visuals can be a valuable and innovative resource for creating patient education materials. Among the systems tested, while each tool had its strengths, Whisk offered a particularly engaging experience due to its additional interactive features, making it particularly well-suited for producing customizable and engaging educational content. Universities and healthcare organizations could integrate these tools into digital learning platforms or educational brochures. As AI technology continues to evolve, thoughtful integration of these resources paired with human oversight has the potential to enhance both patient care and professional training environments.

Declaration of generative AI and AI-assisted technologies in the writing process:
During the preparation of this manuscript, the authors utilized ChatGPT to assist with grammar and stylistic review. Following the use of this tool, the authors carefully reviewed and edited the content as necessary and accept full responsibility for the final version of the work

Keywords: AI generated images, educational resources, health education.