A Retrospective Engineering Analysis of Human-AI Teams Using the Sidekick Principles

The speed of development of Artificial Intelligence (AI) capabilities outpaces that of human systems engineering, particularly for the integration of human-AI teams (HAT) in production-level software. The emergence of frameworks for the principled design and development of hybrid systems offers opportunities to close this gap: They can help identify which common HAT components should be built and matured to accelerate the engineering of such hybrid systems. Leveraging the sidekick principles of human-AI teaming in a review of recent efforts, we identified a prioritized list of system requirements for reusable HAT components. We share these insights with a specific emphasis on the analytical microservices we believe are key to instantiating the sidekick principles.