Normative Multiagent Systems for Scientific Collaboration

Research Problem

Collaboration is governed by norms: expectations that structure how participants share resources, coordinate activities, and uphold shared ethical standards. Large language model (LLM) agents can reason about goals and select actions, but we do not know whether LLM-based agents can reliably incorporate the consequences of norm compliance and violation into goal-directed decision making. We attempt to address this gap.

Research Contribution

In progress

Methodology and Evaluation

In Progress

Technologies

Large language models, Normative multiagent systems, MCP, Python.