Are you comfortable taking complete responsibility when the AI inevitably fails to ensure compliance?