I've been on several working groups for AI ethics and they all go after these topics. But they all fail because these topics are WEIRD (White, Educated, Industrialized, Rich, and Democratic.) Try to dig deeper and you find the word salad doesn't stick.
Instead a simple and workable framework for designing AI can be thought of as follows.
1. Ethical
2. Assurance
Where ethics looks at laws and culture
Where assurance looks at Verification & Validation with a weighting against safety and security.
All those topics mentioned can fit in these three categories but are more usable.
I've been on several working groups for AI ethics and they all go after these topics. But they all fail because these topics are WEIRD (White, Educated, Industrialized, Rich, and Democratic.) Try to dig deeper and you find the word salad doesn't stick.
Instead a simple and workable framework for designing AI can be thought of as follows.
1. Ethical
2. Assurance
Where ethics looks at laws and culture
Where assurance looks at Verification & Validation with a weighting against safety and security.
All those topics mentioned can fit in these three categories but are more usable.