Source: https://arara.code.app.br/en/articles/ai-in-software-development-data-2025
Language: en-US
Published: 2026-10-31T17:30:00-03:00

Guides and comparisons

# AI coding statistics 2025: what the data shows

The 2025 data shows near-universal use of AI in software development, but trust is not keeping up. In the Stack Overflow survey, 84% use or plan to use AI tools, and 46% distrust their accuracy. In Google's DORA report, 90% use AI at work. In the METR study, experienced developers took 19% longer with AI, yet believed it had made them 20% faster.

Arara Code

Published and sources checked on October 31, 2026.

Published by the Arara Code project based on the official documentation cited.

## AI adoption in software development is near-universal

According to the 2025 Stack Overflow Developer Survey, 84% of respondents use or plan to use AI tools in their development process, up from 76% the year before. Among professional developers, 51% use AI tools daily. Google's 2025 DORA report surveyed nearly 5,000 technology professionals, and 90% of them report using AI at work. On its blog, Google adds that they spend a median of two hours a day working with AI.

GitHub sees the same trend. According to its Octoverse 2025 report, 80% of new developers on GitHub use GitHub Copilot in their first week. More than 1.1 million public repositories use an LLM SDK. In August 2025, TypeScript overtook Python and JavaScript to become the most used language on GitHub. GitHub sees this as a shift toward typed languages, which make agent-assisted coding more reliable in production.

[GitHub: Octoverse 2025](https://github.blog/news-insights/octoverse/octoverse-a-new-developer-joins-github-every-second-as-ai-leads-typescript-to-1/)

## Trust is not keeping pace with adoption

The 2025 Stack Overflow survey shows the other side. Positive sentiment toward AI tools fell from over 70% in 2023 and 2024 to 60%. Now 46% distrust the accuracy of these tools, versus 33% who trust it, and only 3% highly trust the output. Experienced developers are the most cautious, with the lowest “highly trust” rate (2.6%) and the highest “highly distrust” rate (20%).

Google's DORA 2025 report also records doubt: 30% of respondents report little or no trust in AI-generated code, slightly fewer than the year before. In the Stack Overflow survey, 87% are concerned about the accuracy of AI agents and 81% about the security and privacy of data. Using is not the same as trusting: people who use AI still need to review and test what it produces.

[Stack Overflow 2025 Developer Survey: AI](https://survey.stackoverflow.co/2025/ai)

Related

[Best AI IDEs in 2026: 6 options and how to chooseBest AI IDEs in 2026: six options, including Cursor, Kiro, Google Antigravity, Zed, and Arara Code, and how to compare agent, model, and cost.](https://arara.code.app.br/en/articles/best-ai-ides)

## Productivity: perception and measurement disagree

In July 2025, METR published a randomized controlled trial with 16 experienced developers from large open-source projects, working on 246 real issues from repositories they already contributed to. With early-2025 AI tools, they took 19% longer. Before the study, they expected a 24% speedup. Afterward, they still believed AI had made them 20% faster. METR notes that the study does not prove AI fails to speed up most developers.

Google's DORA 2025 report completes the picture. More than 80% of respondents believe AI has increased their productivity, and adoption is linked to higher delivery throughput. But it is also linked to lower delivery stability. For DORA, AI amplifies what a team already has, good and bad. Without strong automated testing, mature version control and fast feedback loops, more changes lead to more instability.

[METR: study of experienced developers (2025)](https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/)

## How to use this data when you choose an AI code editor

The METR study shows that the feeling of speed can mislead, so measure on your own code. Pick two or three real tasks from your repository and record the time, the manual fixes and the test results. Keep what DORA points to as the foundation: automated tests, version control and fast feedback. Choose an editor that shows the diff of every change and asks permission before it edits files or runs commands.

In Arara Code, you can compare assistants without switching editors. Download the IDE, create an account on the Free plan, which requires no card, and connect the assistants you already use in the side chat, such as Codex, Claude Code or Kiro. Each one needs its own CLI, login and quota. Create one branch per assistant and ask for the same task. Approve or refuse each edit and command, review the diff, run the tests and compare your notes.

[Google Cloud: 2025 DORA report](https://cloud.google.com/blog/products/ai-machine-learning/announcing-the-2025-dora-report)

## Sources and further reading

- [Stack Overflow 2025 Developer Survey: AI](https://survey.stackoverflow.co/2025/ai)
- [Google Cloud: 2025 DORA report](https://cloud.google.com/blog/products/ai-machine-learning/announcing-the-2025-dora-report)
- [METR: study of experienced developers (2025)](https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/)
- [GitHub: Octoverse 2025](https://github.blog/news-insights/octoverse/octoverse-a-new-developer-joins-github-every-second-as-ai-leads-typescript-to-1/)
- [Google: highlights of the 2025 DORA report](https://blog.google/technology/developers/dora-report-2025)

[Read as Markdown](https://arara.code.app.br/en/articles/ai-in-software-development-data-2025.md)

## Keep exploring

### [Claude Code, Codex and Kiro in one editor: how to use them](https://arara.code.app.br/en/articles/claude-code-codex-kiro-in-one-editor)

See how to use Claude Code, Codex, Kiro and OpenCode in one IDE: official setup, each provider's account, approvals in the side chat and a comparison test.

[Read the guide](https://arara.code.app.br/en/articles/claude-code-codex-kiro-in-one-editor)

### [Arara Code vs GitHub Copilot: editor, agents and plans](https://arara.code.app.br/en/articles/arara-code-vs-github-copilot)

Looking for a GitHub Copilot alternative? Compare it with Arara Code: workflow, agents, Free plans and a hands-on test to decide on your own project.

[Read the guide](https://arara.code.app.br/en/articles/arara-code-vs-github-copilot)

## Make room for your next idea.

Get to know Arara Code in your own project. Start with the installer and review the changes your assistant makes.

Download Arara Code
