OMAHA, TX, UNITED STATES, March 16, 2026 /EINPresswire.com/ — Artificial intelligence (AI) is increasingly being explored as a tool to support medical decision-making, including in dermatology. Image-based AI systems can analyze photographs of skin conditions and suggest possible diagnoses, which could eventually help clinicians identify diseases more quickly. One such system is ChatGPT-4 Vision, which can interpret images and generate diagnostic suggestions. While this technology shows promise, questions remain as to how accurately it can diagnose real-world skin conditions, particularly across different skin tones.
Dermatologic diagnosis often relies heavily on visual assessment. However, the appearance of many skin conditions can vary depending on a patient’s skin tone, which can make accurate diagnosis more challenging. Researchers are increasingly evaluating whether AI tools perform equally well across diverse patient populations. Ensuring that diagnostic systems work reliably for both lighter skin and skin of color is essential to avoid reinforcing existing health disparities and to ensure equitable care.
A new study published in SKIN: The Journal of Cutaneous Medicine®, evaluated the ability of ChatGPT-4 Vision to diagnose common inpatient dermatologic conditions using images alone. Investigators compiled photographs representing the fifteen most common inpatient skin conditions in the United States, including images across skin tones. The AI system was assessed on whether it could correctly identify the condition or include the correct diagnosis among its top three possible diagnoses. The results showed that ChatGPT-4 Vision correctly diagnosed a higher proportion of images depicting White individuals than those depicting skin of color, though the difference was not statistically significant. Overall accuracy remained modest, and even when considering the AI’s top three diagnostic suggestions, performance did not exceed 75% for either group. Accuracy also varied depending on the specific skin condition being evaluated.
These findings highlight both the potential and the current limitations of AI-assisted dermatologic diagnosis. While tools like ChatGPT-4 Vision may eventually help support clinicians, the study suggests that they are not yet reliable enough to be used as stand-alone diagnostic tools for inpatient skin conditions. Continued research, improved training data, and validation across diverse populations will be important steps toward ensuring that AI systems can safely and effectively support patient care.
SKIN: The Journal of Cutaneous Medicine® is a peer-reviewed online medical journal that is the official journal of The National Society for Cutaneous Medicine. The mission of SKIN is to provide an enhanced and accelerated route to disseminate new dermatologic knowledge for all aspects of cutaneous disease.
For more details, please visit www.jofskin.org or contact jofskin@gmail.com.
Erin X. Wei, MD
UNMC
ebarrett@unmc.edu
Legal Disclaimer:
EIN Presswire provides this news content “as is” without warranty of any kind. We do not accept any responsibility or liability
for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information contained in this
article. If you have any complaints or copyright issues related to this article, kindly contact the author above.
