Newshot

Trustworthy AI Teammates for Software Engineering: A Research Agenda for Lifecycle Context, Explainability, and Agentic Collaboration

Summary

arXiv:2611.21624v1 Announce Type: new Abstract: Large Language Models are rapidly changing software engineering (SE) practice by supporting code generation, testing, debugging, documentation, and program comprehension. Yet most current AI coding assistants remain reactive tools: they respond to local prompts, rely on limited context, and provide suggestions whose rationale is often difficult for developers to inspect. This paper argues that the next phase of AI-assisted SE requires a shift from coding assistants to trustworthy AI teammates. We present a research agenda for human-AI co-evolution in SE, organized around three requirements. First, AI systems must become lifecycle-aware, reasoning across requirements and other deployment artifacts rather than isolated files or prompts. Second, they must provide principled explanations that expose assumptions, evidence, trade-offs, and risks in ways that support developer judgment. Third, they must enable agentic collaboration, in which spe…

Original text(via arXiv cs.SE)

Computer Science > Software Engineering

Title:Trustworthy AI Teammates for Software Engineering: A Research Agenda for Lifecycle Context, Explainability, and Agentic Collaboration

Abstract:Large Language Models are rapidly changing software engineering (SE) practice by supporting code generation, testing, debugging, documentation, and program comprehension. Yet most current AI coding assistants remain reactive tools: they respond to local prompts, rely on limited context, and provide suggestions whose rationale is often difficult for developers to inspect. This paper argues that the next phase of AI-assisted SE requires a shift from coding assistants to trustworthy AI teammates. We present a research agenda for human-AI co-evolution in SE, organized around three requirements. First, AI systems must become lifecycle-aware, reasoning across requirements and other deployment artifacts rather than isolated files or prompts. Second, they must provide principled explanations that expose assumptions, evidence, trade-offs, and risks in ways that support developer judgment. Third, they must enable agentic collaboration, in which specialized AI agents participate in software workflows while preserving human oversight and accountability. We propose an evaluation framework that moves beyond isolated functional correctness toward collaborative quality, including contextual alignment, trust calibration, maintenance recovery cost, and traceability. We conclude by outlining open challenges for designing AI systems that improve software quality without weakening human understanding, agency, or responsibility.
Comments:Accepted at the 2026 Pan-African Artificial Intelligence and Smart Systems Conference (PA-AI&SS 2026)
Subjects:Software Engineering (cs.SE)
Cite as:arXiv:2611.21624 [cs.SE]
(or arXiv:2611.21624v1 [cs.SE] for this version)
https://doi.org/10.48550/arXiv.2611.21624 arXiv-issued DOI via DataCite (pending registration)

Submission history

Access Paper:

  • HTML (experimental)

Additional Features

  • Audio Summary

Current browse context:

References & Citations

  • Google Scholar
  • Semantic Scholar

BibTeX formatted citation

Bookmark

Bibliographic and Citation Tools

Code, Data and Media Associated with this Article

Demos

Recommenders and Search Tools

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

The text above was extracted automatically from the source page and may miss images, tables or code blocks. The original is authoritative.

Published by arXiv cs.SE
This site only summarises and aggregates; copyright stays with the original author. Headlines and summaries follow the source — where they differ, the original wins.
Read the original ↗