- In a recent paper (2020), clinical researchers from Stanford proposed the use of such a framework, to think about data and the development of AI; they were thinking about radiology data specifically.
- They concluded that clinical data should be a form of public good, used for the benefit of future patients and that the data should be widely available for the development of knowledge and tools to benefit future patients.
- From this, they drew three main conclusions:
- If the clinical data is really not owned by anyone, those who interact with it then have an obligation to ensure that the data is used for the benefit of future patients in societies.
- This data should be widely shared for research and development, and all the individuals and entities with access to that data, then essentially become stewards of that data and become responsible to carefully safeguard the privacy and to ensure that the data is used for developing knowledge and tools for the good.
- Patient consent wouldn’t necessarily be required before the data is used for secondary purposes, such as AI development and training and testing, as long as there are mechanisms in place to ensure that ethical standards are being followed.
- According to this framework, the authors propose that it would be unethical to sell data to the third parties by granting exclusive access in exchange for monetary or any kind of payments that exceed costs.
Justin Ko, MD. Ethical Considerations in AI. 8th World Congress of Teledermatology, Skin Imaging and AI in Skin diseases – November 2020