A responsible AI system must be resilient to adversarial attacks and data drift. Dawe emphasizes with automated alerts for:

Technology alone cannot fix ethical issues; the organization must change.

For Dawe, “explainability is not just for regulators; it is for users who need to trust the system.” In the enterprise, this means:

RAI must be embedded at every stage of the standard CRISP-DM (Cross-Industry Standard Process for Data Mining) cycle: