Towards autonomous normative multi-agent systems for Human-AI software engineering teams

TL;DR

The paper envisions a future where autonomous SE agents powered by a BDIM-SE cognitive architecture drive core software development activities in concert with humans. It proposes a normative, self-regulating NorMAS-SE framework in which deontic norms and commitments govern agent-human collaboration, enabling scalable, transparent teamwork. The contributions include a BDIM-SE architecture with memory-integrated beliefs, a commitment-based coordination mechanism, and a roadmap for evaluating single- and multi-agent performance alongside human factors. If realized, this approach could dramatically accelerate software delivery while improving compliance and trust in Human-AI teams.

Abstract

This paper envisions a transformative paradigm in software engineering, where Artificial Intelligence, embodied in fully autonomous agents, becomes the primary driver of the core software development activities. We introduce a new class of software engineering agents, empowered by Large Language Models and equipped with beliefs, desires, intentions, and memory to enable human-like reasoning. These agents collaborate with humans and other agents to design, implement, test, and deploy software systems with a level of speed, reliability, and adaptability far beyond the current software development processes. Their coordination and collaboration are governed by norms expressed as deontic modalities - commitments, obligations, prohibitions and permissions - that regulate interactions and ensure regulatory compliance. These innovations establish a scalable, transparent and trustworthy framework for future Human-AI software engineering teams.

Figures (3)

• Figure 1: Conceptual architecture of a BDIM-SE agent
• Figure 2: An example excerpt of BDIM-SE Coding Agent
• Figure 3: An example of coordination plans automatically generated from a commitment