Group Cohesion in Multi-Agent Scenarios as an Emergent Behavior
This work addresses the challenge of simulating realistic social dynamics in AI for applications like social robotics or virtual environments, but it is incremental as it builds on existing psychological frameworks.
The paper tackles the problem of generating human-like social behaviors in multi-agent systems by using the PSI cognitive architecture with intrinsic needs, resulting in emergent group cohesion, altruism towards in-group agents, and adversarial tendencies towards out-group agents, with parameterization showing dramatic effects such as increased in-group cohesion from out-group bias.
In this paper, we elaborate on the design and discuss the results of a multi-agent simulation that we have developed using the PSI cognitive architecture. We demonstrate that imbuing agents with intrinsic needs for group affiliation, certainty and competence will lead to the emergence of social behavior among agents. This behavior expresses itself in altruism toward in-group agents and adversarial tendencies toward out-group agents. Our simulation also shows how parameterization can have dramatic effects on agent behavior. Introducing an out-group bias, for example, not only made agents behave aggressively toward members of the other group, but it also increased in-group cohesion. Similarly, environmental and situational factors facilitated the emergence of outliers: agents from adversarial groups becoming close friends. Overall, this simulation showcases the power of psychological frameworks, in general, and the PSI paradigm, in particular, to bring about human-like behavioral patterns in an emergent fashion.