Article by Amrita Sengupta and Shweta Mohandas: “The rapid integration of artificial intelligence in healthcare settings raises questions about the adequacy of existing data protection frameworks, particularly the reliance on informed consent as the primary mechanism for legitimatising the collection and use of health data for AI model training. This paper examines whether informed consent, as operationalized under India’s Digital Personal Data Protection Act (DPDPA) 2023, can serve as a satisfactory legal and ethical basis for using health data in AI development.
Drawing on the historical evolution of consent from medical research contexts to contemporary digital data protection regimes, this paper demonstrates that consent-based frameworks face structural limitations when applied to AI systems. The analysis reveals a trifecta of consent challenges: patients must consent to medical procedures, to digital health record creation, and implicitly to future AI model training, often without comprehending the scope, purpose, or risks of data reuse.
This paper advances three broad analyses: first, the limitations of informed consent in data protection and operationalisation challenges in healthcare, the dilution of patient consent and autonomy in AI model training, and the role of anonymisation for use of data for AI. Recognizing these limitations, the paper proposes alternative governance frameworks that complement individual consent…(More)”.
