The reality of developers using AI | State of Code Developer Survey
summary
Survey of 1,100+ developers finds 72% daily users and 58% using AI in production, but 96% don’t fully trust AI-generated code while only 48% always review it. GitHub Copilot and ChatGPT top the tool list; developers use about four tools on average and ~35% of usage comes from personal accounts. AI output is more verbose, expanding review workload and tech debt, while enforcing automated guardrails (Sonar) correlates with lower vulnerabilities (‑24%), defects (‑20%) and rework.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 1:09How AI Is Reshaping Coding Workflows
- 2:05Overview of the State of Code Survey
- 3:59Key Finding: Daily AI Adoption Rates
- 5:00Most Common AI Use Cases in Coding
- 5:31Developers’ Trust in AI-Generated Code
- 6:16Most Popular AI Coding Tools Today
- 7:12Personal vs Work Accounts for AI Tools
- 8:49AI’s Impact on Developer Toil
- 10:08Where AI Increases or Reduces Toil
- 12:03Guardrails for Responsible AI Coding
- 13:10ROI Benefits for SonarQube Users
related talks
30mHuman vs AI: How to Ship Secure Code
40mBuilding the Future Developer: Trends in Open Source, Education...
26mAI Agents as Your Co-Developer
57mAccelerate Your Developer Productivity with AI Embrace the Future Now!
15mAI's Impact on Development, Post-Production & Support Journeys
18mUsing AI Across the Dev Lifecycle
19mIf AI Writes the Code, Who Runs Production?
32mBeyond the Hype: A Pragmatic Look at How Developers Actually Use AI Tools