Developer survey shows trust in AI coding tools is falling as usage rises

The landscape of software development is undergoing a significant transformation as AI tools become increasingly integrated into everyday workflows. A recent survey conducted by StackOverflow, encompassing responses from 49,000 professional developers, reveals a paradox: while the use of AI coding tools is on the rise, trust in their accuracy is declining. In 2025, an impressive four out of five developers reported incorporating AI tools into their work processes, reflecting a rapid growth trend in recent years. However, the survey highlighted a troubling statistic: confidence in the reliability of these AI solutions has plummeted from 40% in previous years to just 29% this year. This stark contrast underscores the complex dynamics at play as developers navigate the benefits and challenges of AI technologies like GitHub Copilot and Cursor. Developers largely agree on the potential utility of these tools, yet they are still discovering the most effective ways to leverage them. When asked about their primary frustrations, 45% cited issues with AI solutions that are almost accurate but not entirely so. This particular challenge can lead to subtle bugs and complications that are often difficult to detect, especially for junior developers who may feel overconfident relying on AI-generated suggestions. Moreover, over a third of respondents indicated that some of their visits to Stack Overflow were prompted by problems arising from AI-generated code suggestions. This highlights a crucial aspect of the developer experience: while AI can assist in coding tasks, it can also inadvertently introduce issues that necessitate further troubleshooting. Despite advancements in reasoning-optimized models, the inherent unreliability of predictive technology suggests that these challenges may persist, complicating the integration of AI tools in software development practices.