Comprehension Debt (CD) is the growing gap between what a development team knows about its codebase and what it actually needs to understand in order to maintain and modify that codebase effectively. The term is proposed in Comprehension Debt in GenAI-Assisted Software Engineering Projects, which frames it as a socio-cognitive risk introduced by the adoption of generative AI tools in software development: such tools reduce cognitive load, and the understanding that would otherwise have been built in the course of doing the work is not built. That paper argues the concept is distinct from traditional technical debt, because it resides in the collective cognition of a development team rather than in the codebase itself.
Usage
The proposing study describes four patterns by which Comprehension Debt accumulates, observed in an undergraduate software engineering project: AI-as-black-box code acceptance, in which generated code is taken without being understood; context-mismatch debt; dependency-induced atrophy; and verification-bypass. It also names one mitigating pattern, in which the same tools are used as a comprehension scaffold and the developer builds a deeper understanding of the code through the interaction rather than around it.
Because the debt sits in what a team collectively knows rather than in what it has written, the study’s proposed responses are practices rather than code changes: verification practices, structured retrospectives, and active learning assessments.