Study maps AI-native software engineering's 5x growth and new competency model

Mamdouh Alenezi’s systematic review, published on arXiv in June 2026, examines 48 peer-reviewed publications from leading venues (2016–2026) to understand how Generative AI, LLMs, and Agentic AI are transforming software engineering. The study employs a four-agent research workflow (literature discovery, scientometric analysis, curriculum transformation, workforce impact) and identifies a clear inflection point: annual LLM-for-SE output grew roughly five-fold after late 2022. Key contributions include a conceptual framework organized around intent, collaboration, and verification; a nine-dimension competency model spanning specification, critical evaluation, agent orchestration, and metacognition; and a four-phase university curriculum roadmap with AI-resilient assessment strategies. The paper also outlines faculty-development and workforce-transformation strategies alongside a prioritized agenda of 11 research gaps.

The evidence base is internally contradictory on the magnitude and direction of productivity effects, underscoring that benefits are strongly context-dependent. This suggests that simply adopting AI tools does not guarantee productivity gains—effective use requires judgment, verification, and orchestration skills. The central challenge for the AI-native era is educating engineers for these higher-order competencies rather than code production alone. For professionals, the study signals a fundamental shift: developers must evolve from writing code to managing AI agents, interpreting outputs, and ensuring system integrity. Universities will need to overhaul curricula to produce graduates who can thrive in this new paradigm.