Skin cancers or malignant melanoma have dangerous consequences for UK patients. Primary care physicians fail to distinguish them from benign skin ailments. They frequently refer these patients to dermatology departments. Due to advances in AI, the burden of skin lesion detection can get shifted to computers. This study explores the role of computerized techniques in carrying out diagnostics.

AI-powered software on computers can perform image analyses. Four authors explored the efficacy of AI systems in analyzing skin lesions. Important concepts like convolution neural networks (CNN) and deep learning are the primary methods utilized. Hundreds of skin lesion pictures were the input for the AI system training. AI system got trained with such a large number of such data sets. However, it had a limitation, a bias in favor of the Caucasian skin color. 

The results showed a high degree of accuracy in detecting skin lesions. Multiple AI systems even successfully diagnosed skin cancer or melanoma. The AI systems working on CNN performed as well as the dermatologist, if not better. They could also differentiate between melanoma and benign or harmless skin lesions. However, legal rules hold the clinician responsible rather than the AI system.

The AI systems have skin color bias and may be ineffective. Their efficacy in diagnosing patients with non-white skin color is not evident or proven. They cannot be a replacement for the dermatologists.