State of Code Developer Survey report [pdf] (www.sonarsource.com)

The 2026 State of Code Developer Survey, conducted by Sonar among over 1,100 developers, reveals that AI coding tools have become integral to daily development workflows, with 72% of respondents using them regularly and reporting that 42% of their code is now AI-generated or assisted. However, while developers express a 35% personal productivity boost from AI, a significant trust gap remains; 96% do not fully trust that AI-generated code is functionally correct. This highlights a crucial verification bottleneck, where nearly half of the developers do not routinely check AI-assisted code before committing it, leading to concerns about the reliability of AI outputs. The report underscores the dichotomy in AI's current role in software development. Although developers utilize AI effectively for exploratory tasks like writing documentation and prototyping—where it has shown to be highly effective—their confidence wanes when it comes to modifying complex, existing code. This trend points to an evolving landscape in software engineering, where while AI can enhance efficiency, it also creates new challenges in quality assurance and technical trust that organizations must navigate. As AI continues to influence coding practices, understanding these dynamics will be vital for leveraging AI’s potential without compromising code integrity.