Table 1 Summary of guidance
From: Embedded ethics: a proposal for integrating ethics into the development of medical AI
Domain | Guidance |
|---|---|
Aims | 1. Embedded ethics should anticipate, identify, and address ethical issues that arise during the process of developing medical AI |
2. Embedded ethics should work collaboratively with the development team to consider and address these issues via an iterative and ongoing process | |
Integration | 3. Embedded ethics should involve regular exchanges, formally or informally, between ethicists and technical members of the team |
Practice | 4. Theoretical frameworks employed by embedded ethics should be made clear and explicit |
5. Theoretical frameworks and resulting positions should be explained and justified in terms of specific project goals | |
6. The decision-making structure within the team should be clearly established at the beginning of the process | |
7. Embedded ethics should consider ways that transparency of analyses in the development of the medical AI could be achieved within the restrictions of confidentiality and intellectual property | |
Expertise and training | 8. Embedded ethics calls for expertise in ethical analysis and proficiency in applied settings, as well as a basic understanding of the AI technology in question and its clinical field of deployment |
9. Opportunities should be created, both before and during a project, for participants to acquire the relevant knowledge and skills to do embedded ethics |