Comprehensive Summary
In this study, Nebeker and colleagues developed a definition for "ethically sourced data" by creating four health data repositories and applying supply chain management (SCM) literature to them. The investigation utilized a Value Sensitive Design (VSD) approach to evaluate the ethical health data repository developed and its agreement with SCM. Then. the value-sensitive framework was mapped by investigating ethical principles, including privacy considerations, and synthesizing the resulting insights into a comprehensive set of guidelines. It was found that health data, to be considered ethically-sourced, should be collected considering the following factors: patient consent, cyber security of the data, data bias, and accuracy of the dataset. To ensure ethically principled research, the authors suggest these guidelines should be considered and respected by researchers. Through collecting datasets with ethical guidelines in mind, quality and efficacy can be improved through accounting for bias. However, the paper is limited by its focus on the immediate assembly of the datasets, but not the downstream ethical considerations of the use of said datasets.
Outcomes and Implications
As the use of AI in health research increases, it is important to consider the ethics of the datasets the Large Language Models (LLMs) use to produce results and to ensure transparency and traceability with the public. Due to consumers' placement of importance on ethical considerations in regard to AI, entities will occasionally "ethics wash" products, meaning that ethics are used as a marketing label without actual consideration for them. Therefore, the development of ethical guidelines regulating the collection of health related datasets, and their proper implementation, is critical for the purposes of both consumer protection and ethical research. Additionally, regulations will ensure audibility and accountability on the part of data-collecting efforts. Through leveraging VSD and SCM for ethically-sourced datasets, clinicians will be able to make personalized care plans more efficiently. This will lead to an increased level of trust between patients and healthcare teams who chose to implement AI in their treatment strategies.