Mechanisms should be put in place to ensure responsibility and accountability for AI systems and their outcomes, both before and after their implementation.