Topics:

Survey Reveals Declining Trust in AI Coding Tools Despite Rising Usage Among Developers

Survey shows increased use of AI coding tools among developers but reveals significant trust issues.

Key Points

  • • 80% of developers use AI tools in their workflow
  • • Trust in AI accuracy has fallen from 40% to 29%
  • • 45% of developers struggle with nearly correct AI solutions
  • • Over a third report returning to Stack Overflow for help due to AI issues

A recent survey conducted by Stack Overflow has unveiled a striking paradox in the realm of AI coding tools. While the use of these tools has soared, with 80% of developers incorporating them into their workflows in 2025, trust in their accuracy has plummeted to just 29%, a significant drop from 40% observed in previous years. This decline in confidence is attributed to ongoing challenges in reliability, as around 45% of developers have reported frustration with AI solutions that are nearly correct but often lead to hard-to-detect bugs, especially affecting less experienced developers.

The rising skepticism has revealed a growing reliance on external resources, with over one-third of respondents turning back to Stack Overflow for support due to issues stemming from AI-generated suggestions. Despite these concerns, many developers continue to leverage AI tools, often driven by managerial expectations or the undeniable utility these tools provide. However, the survey emphasizes the urgent need for enhanced training and understanding of effective AI integration in coding practices. Many respondents believe that AI-generated responses should serve as a starting point rather than conclusive solutions, advocating for a collaborative approach similar to pair programming.

Addressing these findings, Jody Bailey, the Chief Product and Technology Officer at Stack Overflow, noted that the platform is committed to improving AI tool literacy within the community to better navigate these challenges. As the AI landscape evolves, it becomes clear that developers require more than just access to tools; they need the skills to apply them effectively and mitigate potential pitfalls in their coding practices.