Home // GREEN 2025, The Tenth International Conference on Green Communications, Computing and Technologies // View article


The Unaccounted Carbon Cost of AI-Assisted Software Engineering: A Hidden Debt and Sustainability Challenge

Authors:
Nelly Nicaise Nyeck Mbialeu
Benjamin Leiding

Keywords: carbon debt; sustainability; technical debt; AI DevTools; green software engineering.

Abstract:
Software engineering is experiencing the impact of AI on its productivity through rapid code generation, code fixes, and workflow automation. However, there is a hidden cost to this convenience, namely a growing double debt of carbon emissions and technical inefficiencies that jeopardise sustainability. Carbon debt is discussed in this paper, referring to the invisible and cumulative environmental damage resulting from the frequent use of AI-driven tools. AI sustainability discussions often overlook the impact of inference phase emissions in this field, where productivity tools lack built-in insights that measure hidden, accumulated environmental burdens. There is a lack of a conceptual and structured method to incentivise carbon debt. This paper conceptually illustrates the negative contribution of AI-assisted development tools, leading to pragmatic mitigation strategies and a preliminary formalization of a measurable sustainability framework for AI-driven development workflows.

Pages: 1 to 7

Copyright: Copyright (c) IARIA, 2025

Publication date: October 26, 2025

Published in: conference

ISSN: 2519-8483

ISBN: 978-1-68558-311-8

Location: Barcelona, Spain

Dates: from October 26, 2025 to October 30, 2025