Agile Ethics for AI
- Values: { explainability fairness accountability privacy }
- Categories: { model-agnostic }
- Stage: { design phase preprocessing in-processing post-processing }
- References:
Butnaru and others associated with the HAI center at Stanford set up a Agile Ethics workflow in the form of a Trello board. From left to right, the workflow walks you through relevant ethical considerations at the various steps of a machine learning pipeline. The phases are:
- Scope
- Consider ethical implications of the project
- Consider skill mapping (what’s the impact of AI on jobs)?
- Facilitates up-skilling or a change of strategy in the use of human talent
- Data audit
- Led by Chief Data Officer
- “Meet and plan” stage in Agile
- Helpful: Data Ethics Canvas
- Train
- Build stage in Agile
- Consider (tools for) transparency and fairness
- Analyse
- Benchmarks, including benchmarks related to e.g. fairness
- Correct e.g. bias where necessary
- Feedback
- Similar to the “review” stage in Agile
- Wizard of Oz experiments to assess acceptance rate prior to deployment
- Potential resource here is the Technology acceptance model (TAM)
- Calibrate
- With a focus on machine-human interaction
- Augment, e.g.
- In which ways does AI augment a job? And which skills cannot and/or should not be replaced?
- In which ways can users augment the AI?
- People & Environment
- Long-term accountability with respect to the impact on people and the environment where AI is deployed.