Introduction. Currently, there is an increase in the incidence of malignant skin tumors. This increases the burden on the healthcare system and necessitates the search for and implementation of new methods for the early diagnosis of skin neoplasms. Among these modern diagnostic methods, artificial intelligence (AI) programs for smartphones (mobile applications) have begun to occupy a significant place. The aim of the study was to compare the accuracy of the preliminary diagnosis of benign and malignant tumors from the epidermal skin tissue between dermatologists and the AI program "Derma Onko Check". Materials and methods. The effectiveness of the Derma Onko Check computer vision program for visual identification and differentiation between benign and malignant tumors from epidermal skin tissue was evaluated in 130 patients aged 24 to 92 years. The conclusions of the program were compared with the final clinical diagnoses of experienced dermatologists and the results of morphological research methods. Results. In the preliminary differential diagnosis of benign and malignant tumors from the epidermal skin tissue, the Derma Onko Check AI program has shown high diagnostic accuracy. The effectiveness of the Derma Onko Check AI program (diagnostic accuracy 86.2%, sensitivity 91.8%, specificity 82.7%) is comparable to that of experienced dermatologists (diagnostic accuracy 92.98%, sensitivity 95.7%, specificity 100%). Conclusion. The computer AI program (for smartphone) "Derma Onko Check" is a simple and effective tool for differential diagnosis between benign and malignant epidermal skin tumors with high diagnostic accuracy at the initial medical examination. The use of the Derma Onko Check program makes it possible to detect malignant epidermal skin tumors early during initial admission by general practitioners and specialist doctors, which will optimize the routing of such patients - referral to a dermatologist or oncologist. This will save health care efforts and resources and speed up cancer diagnosis.