Demystifying the Ethical Framework for Generative AI in Healthcare: A Data Science Perspective

Generative Artificial Intelligence, also known as Generative AI has enhanced the capabilities of what is known to be AI to the world by its capability to generate text, image, or other forms of media through the extensive use of Large Language Models (LLMs). While the roots of LLM can be traced back to the Markovian theories proposed in 1906, Generative AI is the outcome of advancements in transformer-based deep neural networks which extend the machine learning paradigm. Fundamentally, there is a definitive need to store, manage, and use data to generate data that matches consumer needs. Generative AI-based tools today free users from the need to master programming language. This has paved the way for easy access to data generation, analytics, and subsequent dissemination of findings as per the consumer’s needs, often with or without a subscription fee. The ethical frameworks of AI are built on the four key principles, namely, beneficence, non-maleficence, autonomy, and justice. In recent years, explicability, which incorporates intelligibility and accountability, was added as the fifth crucial principle in the framework. The use of AI by organizations have also led to reputational, regulatory, and legal risks, resulting in widespread discussions on Ethical AI or the ethical use of AI.

In India, the Indian Council of Medical Research (ICMR, India) has released guidelines for the Ethical Use of AI in Healthcare. Protecting data privacy at the individual (or citizen) level has been one of the crucial challenges in the healthcare sector. With the advent of Generative AI, these challenges have also experienced a multiplicative effect. In addition, the post-COVID-19 era has led to increased use of digital health technologies, fueling data privacy and security risks apart from misinformation (leading to infodemics) and bias. Such concerns often affect developing countries, especially in the healthcare sector.

The present research provides a state-of-the-art review of the ethical frameworks for Generative AI in healthcare. The study also provides an overview of privacy-preserving Generative AI paradigms, enabling the policy-makers (government, private, and other not-for-profit entities) to plan, propose, and disseminate policies that preserve the privacy of the data shared at an individual level. The study will benefit researchers by developing methodologies that align with the ethical framework for Generative AI, thus aligning with the principles of using AI for Good.