Logo image
Back
Trustworthiness of Vibe Coding from the NIST AI RMF Perspective
Conference proceeding   Open access

Trustworthiness of Vibe Coding from the NIST AI RMF Perspective

Olivia Bunch and Maryam Taeb
Proceedings of the 2026 ACM Southeast Conference, pp.229-234
ACM Other Conferences
ACMSE 2026: 2026 ACM Southeast Conference (Try University, Troy, Alabama, USA, 04/23/2026–04/25/2026)
04/23/2026
DOI: https://doi.org/10.1145/3746467.3801530

Metrics

1 Record Views

Abstract

Generative AI (GenAI) has introduced 'vibe coding', a transformative paradigm where developers use high-level natural language to iteratively generate software. While this approach offers significant gains in development efficiency (with some studies reporting productivity increase of up to 55%), it introduces critical concerns regarding reliability, security and long-term maintainability. This paper analyzes the dual impact of AI coding tools on the software development life cycle (SDLC), identifying recurring vulnerabilities such as insecure code patterns and supply-chain risks. Synthesizing findings from recent studies, this paper assesses how the conversational nature of vibe coding can lead to over-trust and technical debt. Finally, drawing on the NIST AI Risk Management Framework (AI RMF), we propose a governance framework that emphasizes human accountability and phase-dependent controls to ensure the trustworthy adoption of AI-assisted programming. The findings are intended to provide practitioners with a clear framework for governing the use of AI coding tools in their workflows.
url
Trustworthiness of Vibe Coding from the NIST AI RMF PerspectiveView
Published (Version of record) link to article Open CC BY V4.0

Related links

Details

Logo image