Machine Behavior Research on Cooperative AI

Scope of the Workshop

In an era where artificial intelligence (AI) is rapidly transforming the way we live and work, the importance of fostering effective cooperation between humans and AI cannot be overstated. This cooperation is essential for maximizing the potential benefits of AI in enhancing and augmenting human intellect by working together. Indeed, in recent years many efforts have oriented on creating AI systems that are able and willing to work together with humans for specific tasks, with the ultimate goal of creating a synergistic relationship between the strengths of humans and AI systems.

However, current research practices face certain limitations. The literature on human-AI cooperation spans a wide range of tasks, many of which may not inherently support cooperative behavior. Additionally, different domains employ varied conceptual frameworks and experimental settings, leading to a lack of coherence in research approaches. To advance the field, it is vital to identify and exchange current methodologies in human-AI cooperation research. Therefore, the main goals of this workshop are to discuss and address:

• Conceptual Frameworks: What are the current conceptual frameworks being used in human-AI cooperation research? Are there any missing frameworks that should be considered?
• Research Environments: What types of environments (real-life, simulated, controlled) are currently used, and which are the most suitable for studying human-AI cooperation?
• Metrics: What are the state-of-the-art metrics used to quantify human-AI cooperation?
• Relationship Configurations: What configurations of human-AI relationships are studied? Which configurations are missing, and which are the most suitable for research?

Agenda (November 14th - November 15th, 2024)