From Algorithms To Accountability: The Role Of Machine Learning In Skin Cancer Prediction And Its Legal Implications

A. Samiha Marium & Sowbarnika. V, Sastra Deemed University

ABSTRACT

Growing concerns about skin cancer, especially melanoma, are being raised in India and around the world due to changes in lifestyle and greater UV exposure. Skin cancers account for a large share of the estimated 12% increase in cancer incidence expected to occur in India over the forthcoming five years. In order to improve patient outcomes, early identification is essential since prompt surgical treatments can result in effective therapy. Using large datasets including digital health records and picture databases, this study investigates the use of machine learning (ML) & artificial intelligence (AI) to improve the precision and efficacy of skin cancer diagnosis. The study starts with a summary of machine learning (ML) concepts and their particular applications in dermatology, showcasing different algorithms that use data to learn and enhance diagnostic performance. The variety of skin malignancies makes it difficult to create a uniform classification model, even with advances in technology, especially when it comes to precisely separating lesions from tissue surrounding them. This is crucial for a successful diagnosis because uneven lesion borders can substantially affect the accuracy of predictions. In addition, the study discusses the regulatory ramifications of AI-powered skin cancer detection technologies as well as the challenges associated with assigning blame when ML systems are involved in medical blunders. By looking at these elements, the study hopes to shed light on how to successfully incorporate machine learning (ML) into the prediction of skin diseases, highlighting the need of responsibility and safety when using AI-powered medical devices to diagnose skin cancer.

Keywords: Machine learning (ML), Convolutional Neural Networks (CNN), Consumer Protection Act (CPA), Medical Devices Rules, 2017.