Keeping an Eye on AI: A Framework for Effective Human Oversight of AI Systems

The researchers propose a practical framework for effective human oversight of AI systems, integrating interdisciplinary perspectives from computer science, human-computer interaction, psychology, philosophy, and law. The framework provides a working definition, architectures, and processes for human oversight, and offers initial templates for documenting these architectures and processes, which have been applied in multiple domains. The authors also outline open research challenges in this emerging field.