23 Introduction: Artificial intelligence (AI) in healthcare has gained momentum with advances in 24 affordable technology that has potential to help in diagnostics, predictive healthcare and 25 personalized medicine. In pursuit of applying universal non-biased AI in healthcare, it is 26 essential that data from different settings (gender, age and ethnicity) is represented. We present 27 findings from beta-testing an AI-powered dermatological algorithm called Skin Image Search, by 28 online dermatology company First Derm on Fitzpatrick 6 skin type (dark skin) dermatological 29 conditions. Methods: 123 dermatological images selected from a total of 173 images 30 retrospectively extracted from the electronic database of a Ugandan telehealth company, The 31 Medical Concierge Group (TMCG) after getting their consent. Details of age, gender and 32 dermatological clinical diagnosis were analyzed using R on R studio software to assess the 33 diagnostic accuracy of the AI app along disease diagnosis and body part. Predictability levels of 34 the AI app was graded on a scale of 0 to 5, where 0-no prediction made and 1-5 demonstrating 35 reducing correct prediction. Results: 76 (62%) of the dermatological images were from females 36 and 47 (38%) from males. The 5 most reported body parts were; genitals (20%), trunk (20%), 37 lower limb (14.6%), face (12%) and upper limb (12%) with the AI app predicting a diagnosis in 38 62% of image body parts uploaded. Overall diagnostic accuracy of the AI app was low at 17% 39 (21 out of 123 predictable images) with varying predictability levels correctness i.e. 1-8.9%, 2-40 2.4%, 3-2.4%, 4-1.6%, 5-1.6% with performance along individual diagnosis highest with 41 dermatitis (80%). Conclusion: There is a need for diversity in the image datasets used when 42 training dermatology algorithms for AI applications in clinical decision support as a means to 43 increase accuracy and thus offer correct treatment across skin types and geographies.
44