TLDR: The 2025 Stack Overflow Developer Survey reveals that 23% of developers are regularly using AI agents in their work, despite significant concerns regarding the accuracy of AI outputs (87%) and data privacy/security (81%). While 70% of users report productivity gains, challenges like ‘almost right’ AI solutions and high platform costs persist.
The 2025 Stack Overflow Developer Survey, a definitive report on the state of software development, indicates a growing adoption of AI agents among developers. According to the survey, 23% of developers are regularly utilizing AI agents in their professional work on a weekly basis, with an additional 8% using them infrequently. Furthermore, 17% of developers are planning to integrate AI agents into their workflows, highlighting a clear trend towards autonomous software entities.
Despite this increasing adoption, the survey uncovers substantial concerns among the developer community. A significant 87% of developers expressed worry about the accuracy of information generated by AI agents. Similarly, 81% are apprehensive about the privacy and security of data when interacting with these tools. The financial aspect also presents a barrier, with 53% of respondents citing the cost of AI agent platforms as a concern.
However, the perceived benefits are a driving force behind the adoption. Approximately 70% of developers who have used AI agents believe these tools have successfully reduced the time spent on specific development tasks. A comparable figure, 69%, at least somewhat agree that AI agents have led to overall productivity gains.
Conversely, AI agents have shown less impact in certain areas. Only 17% of developers believe these tools have improved team collaboration, and a mere 36% feel they have enhanced code quality. A notable frustration, cited by 45% of respondents, is dealing with ‘AI solutions that are almost right, but not quite,’ which often leads to more time spent on debugging. In fact, 66% of developers report spending more time fixing ‘almost-right’ AI-generated code.
Also Read:
- Legit Security Enhances AI-Developed Code Security with New Agent-Powered Capabilities
- Federal Agencies See Ninefold Surge in Generative AI Adoption, GAO Reports
Taking a broader view of AI in development, the survey found that 78% of respondents use AI tools in their development process, a significant increase from 62% in the previous year’s study. Despite this widespread usage, 46% of developers do not trust the accuracy of AI tool output, while only 33% do. Nevertheless, 59% maintain a favorable inclination towards using AI tools in their workflow, compared to 21% with an unfavorable view. The 2025 survey, which gathered over 49,000 responses from 177 countries, included a new focus on AI agent tools, LLMs, and community platforms, providing comprehensive insights into the evolving landscape of software development.
