Reproduce “Facilitating Trustworthy Human-Agent Collaboration in LLM-based Multi-Agent System oriented Software Engineering” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Krishna Ronanki
Krishna Ronanki from Chalmers University of Technology and the University of Gothenburg proposes a RACI-based framework to guide task allocation and foster trustworthiness in human-agent collaboration within LLM-based multi-agent software engineering systems. This framework aims to ensure human oversight and align with emerging Trustworthy AI principles, demonstrated through a DevOps planning phase example.
Key insights
- Established project management tools, like the RACI matrix, can be effectively adapted to structure and clarify responsibilities in novel human-AI multi-agent collaborative environments.
- Trustworthy AI principles and regulatory requirements can be operationalized and embedded directly into software development workflows through a structured approach to responsibility assignment.
- A clearly defined framework allows for an optimal synergy, leveraging human strengths in critical judgment and creativity alongside AI capabilities in automation and data processing, while maintaining accountability.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.