EASE© framework in design and development of clinical artificial intelligence applications

Building Clinical artificial intelligence (AI) applications requires a delicate balance between clinical need, technical knowhow and ethical considerations. Many Clinical AI models have issues like unproven value, model opacity, model drift, disutility and persistent resistance for adoption. At Apollo Hospitals, we design, develop, validate and deploy Clinical AI solutions with ethically sourced clinical data which feature facets from accountability and accuracy to fairness and inclusivity. Without proper oversight, Clinical AI applications can quickly become flawed, hindering their potential impact. To prevent such issues, the EASE framework was developed to guide the processes towards the right path. EASE encompasses Ethics, Adoption, Suitability and Explainability. As digital technology continues to permeate all aspects of society, the discussion and debate surrounding data integrity will only grow more pressing. Data ethics are relevant to everyone—government, businesses, and individuals alike—and should be approached with openness and transparency. The process of discussion and knowledge-sharing is vital for raising awareness and promoting ethical practices.