International Journal of Science, Engineering and TechnologyAuthors: Saanvi Gupta
Abstract: As of 2026, the integration of Artificial Intelligence (AI) into the Software Development Life Cycle (SDLC) has shifted from assistive autocomplete to autonomous "Agentic AI." This paper investigates the "Productivity-Quality Paradox," examining how the transition to AI-driven development impacts long-term system sustainability, architectural integrity, and technical debt. The study synthesizes longitudinal data from 2020–2025, available on open access databases and formed the structure of this research article. Results indicate a significant bimodal effect. While AI accelerates Minimum Viable Product (MVP) development by 40–60% and improves automated test case accuracy to nearly 98%, it has simultaneously triggered a sustainability crisis. Key metrics reveal a 4x increase in code duplication (violating DRY principles) and a doubling of code churn compared to 2021 baselines. Furthermore, a "Verification Bottleneck" has emerged: despite perceived speed gains, experienced developers spend 19% more time "chaperoning" and debugging AI-generated logic. Security remains a critical failure point, with over 51% of AI-authored code containing vulnerabilities. The research introduces the concept of "Agentic Debt"—the hidden cost of autonomous, repository-wide modifications without human contextual oversight. To mitigate systemic decay, the paper proposes a transition to the SPACE productivity framework and the implementation of AI-aware CI/CD pipelines. The study concludes that while AI is an unmatched force multiplier for speed, human-in-the-loop (HITL) verification remains the only safeguard against long-term technical bankruptcy.
DOI: https://doi.org/10.5281/zenodo.18359795
