TL;DR
Human-AI collaboration is often poorly defined, leading to confusion in application. This study categorizes human-AI teams into distinct types based on their interaction dynamics and roles.
✦ Why It Matters
Engineers can leverage these team categories to design AI systems that enhance collaboration and efficiency in real-world applications.
Key Takeaways
Full Summary
Human-AI teaming is an emerging area of research, but the variety of studies complicates understanding the types of teams involved. This study analyzes 53 papers and identifies five clusters of human-AI teams based on psychological taxonomies: AI Assistant, Ad-hoc Dependency, Ad-hoc Forced Dependency, Paired Equanimity, and Group Equanimity.
Each cluster reflects unique team characteristics, suggesting that insights from one type may not apply to another. The research emphasizes the need for clearer definitions and reporting standards in human-AI team studies.
A checklist for researchers is provided to help identify and categorize the types of teams being studied. This work aims to synthesize the field and improve the clarity of human-AI team research.
Related