Stack Overflow Developer Survey 2025

3 AI

In this section we gain insight into the real sentiments behind the surge in AI popularity. Is it making a real impact in the way developers work or is it all hype?

3.1 Sentiment and usage

AI tools in the development process

84% of respondents are using or planning to use AI tools in their development process, an increase over last year (76%). This year we can see 51% of professional developers use AI tools daily.

Do you currently use AI tools in your development process?
All RespondentsProfessional DevelopersLearning to CodeEarly Career DevsMid Career DevsExperienced Devs
All RespondentsYes, I use AI tools daily47.1%Yes, I use AI tools weekly17.7%Yes, I use AI tools monthly or infrequently13.7%No, but I plan to soon5.3%No, and I don't plan to16.2%
Responses: 33,662(68.7%)

AI tool sentiment

Conversely to usage, positive sentiment for AI tools has decreased in 2025: 70%+ in 2023 and 2024 to just 60% this year. Professionals show a higher overall favorable sentiment (61%) than those learning to code (53%).

How favorable is your stance on using AI tools as part of your development workflow?
All RespondentsProfessional DevelopersLearning to CodeEarly Career DevsMid Career DevsExperienced Devs
All RespondentsVery favorable22.9%Favorable36.8%Indifferent17.6%Unsure2.3%Unfavorable10.8%Very unfavorable9.6%
Responses: 33,412(68.2%)

3.2 Developer tools

Accuracy of AI tools

More developers actively distrust the accuracy of AI tools (46%) than trust it (33%), and only a fraction (3%) report "highly trusting" the output. Experienced developers are the most cautious, with the lowest "highly trust" rate (2.6%) and the highest "highly distrust" rate (20%), indicating a widespread need for human verification for those in roles with accountability.

How much do you trust the accuracy of the output from AI tools as part of your development workflow?
All RespondentsProfessional DevelopersLearning to CodeEarly Career DevsMid Career DevsExperienced Devs
All RespondentsHighly trust3.1%Somewhat trust29.6%Somewhat distrust26.1%Highly distrust19.6%
Responses: 33,244(67.8%)

AI tools' ability to handle complex tasks

In 2024, 35% of professional developers already believed that AI tools struggled with complex tasks. This year, that number has dropped to 29% among professional developers and is consistent amongst experience levels. Complex tasks carry too much risk to spend extra time proving out the efficacy of AI tools.

How well do the AI tools you use in your development workflow handle complex tasks?
All RespondentsProfessional DevelopersLearning to CodeEarly Career DevsMid Career DevsExperienced Devs
All RespondentsVery well at handling complex tasks4.4%Good, but not great at handling complex tasks25.2%Neither good or bad at handling complex tasks14.1%Bad at handling complex tasks22%Very poor at handling complex tasks17.6%I don't use AI tools for complex tasks / I don't know16.8%
Responses: 33,230(67.8%)

AI in the development workflow

Developers show the most resistance to using AI for high-responsibility, systemic tasks like Deployment and monitoring (76% don't plan to) and Project planning (69% don't plan to).

Which parts of your development workflow are you currently integrating into AI or using AI tools to accomplish or plan to use AI to accomplish over the next 3 - 5 years? Please select one for each scenario.
Currently Mostly AICurrently Partially AIPlan to Partially Use AIPlan to Mostly Use AIDon't Plan to Use AI for This Task
Currently Mostly AISearch for answers54.1%Generating content or synthetic data35.8%Learning new concepts or technologies33.1%Documenting code30.8%Creating or maintaining documentation24.8%Learning about a codebase20.8%Debugging or fixing code20.7%Testing code17.9%Writing code16.9%Predictive analytics11%Project planning10.8%Committing and reviewing code10.2%Deployment and monitoring6.2%
Responses: 11,202(22.9%)

AI workflow and tool satisfaction

Respondents who said they are currently using mostly AI tools to complete tasks in the development workflow are highly satisfied with and frequently using AI to search for answers or learn new concepts; respondents plan to mostly use AI in the future for documentation and testing tasks and are slightly less satisfied with the tools they are using now.

How favorable is your stance on using AI tools as part of your development workflow and which parts of your development workflow are you currently integrating into AI or using AI tools to accomplish or plan to use AI to accomplish over the next 3 - 5 years? Please select one for each scenario.
Currently mostly AICurrently partially AIPlan to partially use AIPlan to mostly use AIDon't plan to use AI for this task
Currently mostly AINumber of responses6,053685Average AI Sentiment Recoded (1 - Very Unfavorable to 6 - Very Favorable)Percent of respondents 5.25 5.3 5.35 5.4 5.45 5.5 5.55 5.6 5.65% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55Commit/ReviewDocsDebug/fixOpsDocumenting codeContent/DataLeaning codebaseLearning techPredictive analyticsProject planningAnswersTesting codeWriting code
Responses: 11,184(22.8%)

AI tool frustrations

The biggest single frustration, cited by 66% of developers, is dealing with "AI solutions that are almost right, but not quite," which often leads to the second-biggest frustration: "Debugging AI-generated code is more time-consuming" (45%)

When using AI tools, which of the following problems or frustrations have you encountered? Select all that apply.
All RespondentsAI solutions that are almost right, but not quite66%Debugging AI-generated code is more time-consuming45.2%I don’t use AI tools regularly23.5%I’ve become less confident in my own problem-solving20%It’s hard to understand how or why the code works16.3%Other (write in):11.6%I haven’t encountered any problems4%
Responses: 31,476(64.2%)

AI and humans in the future

In a future with advanced AI, the #1 reason developers would still ask a person for help is "When I don’t trust AI’s answers" (75%). This positions human developers as the ultimate arbiters of quality and correctness.

In the future, if AI can do most coding tasks, in which situations would you still want to ask another person for help? Select all that apply.
All RespondentsWhen I don’t trust AI’s answers75.3%When I have ethical or security concerns about code61.7%When I want to fully understand something61.3%When I want to learn best practices58.1%When I’m stuck and can’t explain the problem54.6%When I need help fixing complex or unfamiliar code49.8%When I want to compare different solutions44.1%When I need quick help troubleshooting27.5%Other6.1%I don’t think I’ll need help from people anymore4.3%
Responses: 29,163(59.5%)

Vibe coding

Most respondents are not vibe coding (72%), and an additional 5% are emphatic it not being part of their development workflow.

In your own words, is "vibe coding" part of your professional development work? For this question, we define vibe coding according to the Wikipedia definition, the process of generating software from LLM prompts.
All Respondents18-24 years old25-34 years old35-44 years old45-54 years old55-64 years old
All RespondentsYes, emphatically0.4%Yes11.9%Yes, somewhat2.8%I have tried it2.1%Not sure1.2%No72.2%No, emphatically5.3%Uncategorized4%
Responses: 26,564(54.2%)

3.3 AI Agents

AI agents

AI agents are not yet mainstream. A majority of developers (52%) either don't use agents or stick to simpler AI tools, and a significant portion (38%) have no plans to adopt them.

Are you using AI agents in your work (development or otherwise)? AI agents refer to autonomous software entities that can operate with minimal to no direct human intervention using artificial intelligence techniques.
All RespondentsProfessional DevelopersLearning to CodeProfessional AI UsersLearning AI Users
All RespondentsYes, I use AI agents at work daily14.1%Yes, I use AI agents at work weekly9%Yes, I use AI agents at work monthly or infrequently7.8%No, but I plan to17.4%No, I use AI exclusively in copilot/autocomplete mode13.8%No, and I don't plan to37.9%
Responses: 31,877(65%)

AI agents affect on work productivity

52% of developers agree that AI tools and/or AI agents have had a positive effect on their productivity.

Have AI tools or AI agents changed how you complete development work in the past year?
All RespondentsYes, to a great extent16.3%Yes, somewhat35.3%Not at all or minimally41.4%No, but my development work has significantly changed due to non-AI factors2.6%No, but my development work has changed somewhat due to non-AI factors4.5%
Responses: 31,636(64.5%)

AI agent uses at work

If you happen to be using AI agents at work and you are a software developer, chances are high that you are using agents for software development (84%).

What industry purposes or specific tasks are you using AI agents in your development work? Select all that apply from both lists.
Industry PurposeSoftware engineering83.5%Data and analytics24.9%IT operations18%Business process automation17.6%Decision intelligence11.3%Customer service support11.2%Marketing8.6%Cybersecurity7.4%Robotics3.9%Other2.2%
Responses: 12,301(25.1%)

AI agent uses for general purposes

TL;DR: Agents used outside of work are mostly used for language processing tasks (49%).

What industry purposes or specific tasks are you using AI agents in your development work? Select all that apply from both lists.
General PurposeLanguage processing49%Integration with external agents and APIs38.3%MCP servers34.4%Agent/multi-agent orchestration28.1%Vector databases for AI applications24.1%Multi-platform search enablement19.4%Personalized agent creation18.3%Other3%
Responses: 5,797(11.8%)

Impacts of AI agents

The most recognized impacts are personal efficiency gains, and not team-wide impact. Approximately 70% of agent users agree that agents have reduced the time spent on specific development tasks, and 69% agree they have increased productivity. Only 17% of users agree that agents have improved collaboration within their team, making it the lowest-rated impact by a wide margin.

To what extent do you agree with the following statements regarding the impact of AI agents on your work as a developer?
All Respondents27.3%35.9%21.3%8.2%7.3%AI agents have accelerated my learning about new technologies or codebases.29.3%34.9%22.4%7%6.4%AI agents have helped me automate repetitive tasks.17.1%31.9%25.3%14.2%11.5%AI agents have helped me solve complex problems more effectively.6.6%10.7%40.5%20%22.2%AI agents have improved collaboration within my team.12.2%25.3%32.4%17.1%13.1%AI agents have improved the quality of my code.27.7%41%20.4%6%4.9%AI agents have increased my productivity.29.3%40.8%17.8%6.9%5.1%AI agents have reduced the time spent on specific development tasks.Strongly agreeSomewhat agreeNeutralSomewhat disagreeStrongly disagree
Responses: 12,823(26.2%)

Challenges with AI agents

Is it a learning curve, or is the tech not there yet? 87% of all respondents agree they are concerned about the accuracy, and 81% agree they have concerns about the security and privacy of data.

To what extent do you agree with the following statements regarding AI agents?
All Respondents57.1%29.8%9.7%2.3%1.1%I am concerned about the accuracy of the information provided by AI agents.56.1%25.3%11.7%4.7%2.2%I have concerns about the security and privacy of data when using AI agents.16.5%29.7%37.3%12.6%3.9%Integrating AI agents with my existing tools and workflows can be difficult.15.5%27.9%31.8%17.8%6.9%It takes significant time and effort to learn how to use AI agents effectively.13.8%14.4%30.6%15%26.2%My company's IT and/or InfoSec teams have strict rules that do not allow me to use AI agent tools or platforms25.4%27.9%31.8%10.3%4.6%The cost of using certain AI agent platforms is a barrier.Strongly agreeSomewhat agreeNeutralSomewhat disagreeStrongly disagree
Responses: 28,930(59%)

AI Agent data storage tools

When it comes to data management for agents, traditional, developer-friendly tools like Redis (43%) are being repurposed for AI, alongside emerging vector-native databases like ChromaDB (20%) and pgvector (18%).

You indicated you use or develop AI agents as part of your development work. Have you used any of the following tools for AI agent memory or data management in the past year?
All RespondentsRedis42.9%GitHub MCP Server42.8%supabase20.9%ChromaDB19.7%pgvector17.9%Neo4j12.3%Pinecone11.2%Qdrant8.2%Milvus5.2%Fireproof5%LangMem4.8%Weaviate4.5%LanceDB4.4%mem04%Zep2.8%Letta2.5%
Responses: 3,398(6.9%)

AI Agent orchestration tools

The agent orchestration space is currently led by open-source tools. Among developers building agents, Ollama (51%) and LangChain (33%) are the most-used frameworks.

You indicated you use or develop AI agents as part of your development work. Have you used any of the following tools for AI agent orchestration or agent frameworks in the past year?
All RespondentsOllama51.1%LangChain32.9%LangGraph16.2%Vertex AI15.1%Amazon Bedrock Agents14.5%OpenRouter13.4%Llama Index13.3%AutoGen (Microsoft)12%Zapier11.8%CrewAI7.5%Semantic Kernel6%IBM watsonx.ai5.7%Haystack4.4%Smolagents3.7%Agno3.4%phidata2.1%Smol-AGI1.9%Martian1.7%lyzr1.5%
Responses: 3,758(7.7%)

AI Agent observability and security

Developers are primarily adapting their existing, traditional monitoring tools for this new task, rather than adopting new, AI-native solutions. The most used tools for AI agent observability are staples of the DevOps and application monitoring world: Grafana + Prometheus are used by 43% of agent developers, and Sentry is used by 32%.

You indicated you use or develop AI agents as part of your development work. Have you used any of the following tools for AI agent observability, monitoring or security in the past year?
All RespondentsGrafana + Prometheus43%Sentry31.8%Snyk18.2%New Relic13%LangSmith12.5%Honeycomb8.8%Langfuse8.8%Wiz6.9%Galileo6.2%Adversarial Robustness Toolbox (ART)5.5%Protect AI5%Vectra AI4.4%arize3.7%helicone3.2%Metero2.7%opik2.3%
Responses: 2,689(5.5%)

AI Agent out-of-the-box tools

ChatGPT (82%) and GitHub Copilot (68%) are the clear market leaders, serving as the primary entry point for most developers using out-of-the-box AI assistance.

You indicated you use or develop AI agents as part of your development work. Have you used any of the following out-of-the-box agents, copilots or assistants?
All RespondentsChatGPT81.7%GitHub Copilot67.9%Google Gemini47.4%Claude Code40.8%Microsoft Copilot31.3%Perplexity16.2%v0.dev9.1%Bolt.new6.5%Lovable.dev5.7%AgentGPT5%Tabnine5%Replit5%Auto-GPT4.7%Amazon Codewhisperer3.9%Blackbox AI3.5%Roo code (Roo-Cline)3.4%Cody3%Devin AI2.7%Glean (Enterprise Agents)1.3%OpenHands (formerly OpenDevin)1%
Responses: 8,323(17%)
Previous
Technology
Next
Work