Technical Debt in the AI Era

Rapid advances in frontier models and development environments continually reshape technical debt challenges, eliminating some limitations while creating new ones. Regardless of such rapid change, the need to understand and manage Technical Debt will not change. In this paper, we introduce the Technical Debt Manifesto and advocate for following its principles to manage Technical Debt in the era of AI. We discuss values, beliefs, and principles relevant to AI where creation of software will have heterogeneous participants, ranging from humans to different classes of tools and workflows driving different forms of collaboration. The values, beliefs, and principles share a common goal to shift the focus from automation-focused Technical Debt management to human-driven decision-making throughout the development lifecycle, albeit supported by automation.