AI Ethics Council Report: Explainability for AI Predictions

This quarter's AI Ethics Council meeting focused on the responsibility that AI companies have to ensure that AI predications can easily be explained in order to build trust and adoption by decision makers.

Read this report to hear latest recommendations from esteemed council members on how to:

• Enable case-based reasoning
• Cross-validate and calibrate using rigorously defined confidence intervals
• Track data provenance

While explainability still has a long way to go, AI companies can still provide the “tools to uncover” potential rationale, empowering people to either build trust in the systems or find options for recourse.

‍