Developers Lean on AI More, But Report Growing Doubts About Accuracy, Stack Overflow Survey Says

Software developers are using AI tools more than ever, but many say their confidence in the results is slipping, according to Stack Overflow’s 2025 Developer Survey, which tracked rising adoption alongside growing demands for verification.

The survey, fielded from May 29 to June 23, 2025, analyzed 49,009 responses from 166 countries. Developers reported widespread use of AI assistants in programming workflows, while also describing a persistent problem: AI outputs that look plausible, compile, and still waste time.

Stack Overflow said 84% of respondents are using or planning to use AI tools in their development process, up from 76% a year earlier. In the AI section, 47.1% of respondents said they use AI tools daily, and another 17.7% use them weekly.

Yet developers’ trust has not kept pace. Stack Overflow’s analysis said trust in AI accuracy fell to 29% from 40% in prior years, and “positive favorability” dropped to 60% from 72% year over year.

In the survey’s AI breakdown, Stack Overflow summarized the mood starkly: “More developers actively distrust the accuracy of AI tools (46%) than trust it (33%),” with only 3% saying they highly trust AI output.

"Almost right" is still the dominant complaint
Developers’ most common frustration was not that AI cannot write code at all, but that it produces solutions that are close enough to be tempting and wrong enough to be costly.

In the AI section, 66% of respondents selected “AI solutions that are almost right, but not quite” as a problem they encounter. Another 45.2% said debugging AI-generated code is more time-consuming. Stack Overflow’s blog post interpreting the results pointed to the same dynamic, arguing that the next phase of AI in software will be shaped by trust and review rather than novelty.

That caution shows up in how developers imagine future workflows. In a question about a future where AI can do most coding tasks, Stack Overflow said the top reason developers would still ask another person for help was “When I don’t trust AI’s answers” (75%).

Agents are arriving, but not yet standard
The survey also distinguished everyday copilots from more autonomous “AI agents” capable of carrying out multi-step tasks. Daily use of agents at work remains limited: 14.1% of respondents said they use AI agents daily, while 37.9% said they do not plan to use them.

A further 13.8% said they use AI exclusively in copilot or autocomplete mode, suggesting many developers still prefer AI that proposes code inside familiar tooling rather than acting independently across systems.

Even among agent users, skepticism is prominent. Stack Overflow reported that 87% of respondents agreed they are concerned about the accuracy of information provided by AI agents, and 81% agreed they have concerns about security and data privacy when using them.

Which models and tools developers say they use
When asked about specific large language models, developers reported the highest usage for OpenAI’s GPT models. The survey showed 81.4% of respondents said they used OpenAI GPT models for development work in the past year.

In “admired” rankings, Claude Sonnet stood out. Stack Overflow’s results page said Claude Sonnet was the most admired AI model, with 67.5% “admired” among respondents who evaluated that category.

Despite constant churn in developer tooling, mainstream environments remained dominant. Visual Studio Code was used by 75.9% of respondents, and Stack Overflow noted that subscription AI IDEs did not dislodge Visual Studio and VS Code from the top slots.

Programming languages: stability at the top, growth in Python
In languages, JavaScript remained the most used, cited by 66% of respondents, followed by HTML/CSS (61.9%), SQL (58.6%), and Python (57.9%).

Stack Overflow said Python’s adoption “accelerated significantly,” rising 7 percentage points from 2024 to 2025, and linked that growth to its role in AI, data science, and back-end development.

A workflow shift, not a trust shortcut
The pattern running through the survey is that developers are incorporating AI into routine work, but treating it less like an oracle and more like a fast, fallible collaborator. AI tools are widely used for drafting and iteration, while verification and debugging remain developer-owned responsibilities, especially as tasks become more complex.

Stack Overflow’s methodology notes also underscore an important constraint on interpretation: respondents were recruited primarily through Stack Overflow’s own channels, meaning highly engaged users were more likely to see and respond to the survey prompts.

Still, the overall message is consistent across the survey and Stack Overflow’s own write-up: developer AI use is rising, but “trust but verify” is not a slogan so much as the default operating mode.

Posted by John K. Waters on January 20, 2026