Stack Overflow's 2025 survey found 84 percent of developers using or planning to use AI tools, while more distrust their accuracy than trust it. The biggest complaint: answers that are almost right.
AI coding tools have gone from novelty to default in about two years. Stack Overflow's 2025 Developer Survey shows how far adoption has run ahead of confidence.
Nearly everyone is using it
Eighty-four percent of respondents said they use or plan to use AI tools in their development work, up from 76 percent the year before. Forty-seven percent use them every day, rising to 51 percent of professional developers.
Fewer trust it
Trust has not kept pace. About a third of developers said they trust the accuracy of AI output, while 46 percent said they distrust it. Positive sentiment towards AI tools fell to about 60 percent, from above 70 percent in both 2023 and 2024.
Almost right is the problem
The most common frustration, cited by 66 percent, was AI answers that are almost right, but not quite. The second, at 45 percent, was that debugging AI-generated code takes longer. Code that is obviously wrong gets thrown away; code that is subtly wrong gets shipped, and found later.
What it means for the software you commission
- Expect AI in the workflow of any team you hire. Ask how its output is reviewed, not whether it is used.
- Insist on tests. They catch the almost-right code that a quick read misses.
- Budget for review time. Faster first drafts move effort into checking, not away from it.
- Value judgement over volume. The scarce skill is knowing when the plausible answer is wrong.
The developers closest to these tools are among the most sceptical of them. That is not a reason to avoid AI; it is a description of how to use it well.
Sources
Written by Aevion Labs from the sources above. Figures are as reported by each source on the dates shown.