AI-Assisted Coding: Evolution or Erosion of Software Development Skills?

Abstract

In the era of generative artificial intelligence, tools such as GitHub Copilot and ChatGPT are profoundly transforming software development practices. This technological advancement, while facilitating the automation of complex tasks and improving productivity, raises a crucial question: does AI-assisted coding represent an evolution or an erosion of programming skills? This article explores the cognitive, pedagogical, and professional effects of AI assistance in software development. Through a structured analysis, it highlights the proven benefits of AI (speed, error reduction, work comfort), while highlighting the risks associated with non-critical use, including the reduction of algorithmic reasoning, cognitive dependence, and loss of creativity. The study shows that the role of the developer is evolving: they are becoming supervisors, critics, and strategists, rather than mere executors. To support this shift, the article suggests concrete avenues: rethinking training curricula, training in prompt engineering, establishing a culture of human-machine collaboration, and promoting an ethic of increased responsibility.