Surveys of AI’s Impact on Software Development in Recent Years (Category AI)

Surveys are a common way to assess an industry and can provide valuable data, particularly with large samples. Since I am interested in AI’s effect on development, I decided to look at the surveys already conducted, their findings and their methods. I will examine interesting points in later posts; for now, here is the collection.

1. Stack Overflow Developer Surveys (2023, 2024) These are among the largest surveys, covering tens of thousands of engineers who still use Stack Overflow rather than switching entirely to ChatGPT. Both 2023 and 2024 included dedicated AI sections. In 2023, 70% of 90,000 developers used or planned to use AI tools at work. By 2024, that figure was 76%. The leading tools were ChatGPT (83%) and GitHub Copilot (56%). Similar questions across years make these surveys useful for tracking trends. The 2025 survey has already closed, but its results are not yet published. When they appear, I will explore the data and trends.

2. GitHub Developer Survey (August 2024) GitHub surveyed 2,000 developers in the United States, Brazil, India and Germany. It found that 97% used AI tools at work. Depending on the country, 61–73% believed AI improved their ability to build software. There is also an earlier GitHub study from September 2022. Its findings are interesting historically and methodologically, but the information is less relevant to the present situation.

3. JetBrains Developer Ecosystem Survey (2023, 2024) JetBrains has run its survey since 2017 and first included AI questions in 2023. In 2024, 84% of developers were familiar with generative AI tools. Developers were most interested in delegating routine tasks to AI.

4. DORA Research (2023, 2024) The well-known DORA metrics came from the DevOps Research and Assessment team, now part of Google. It publishes annual reports and in 2025 published a separate “Impact of Gen AI in Software Development” report. It found that 76% of developers use AI for everyday tasks, alongside a paradox: an increase of 2.1% in individual productivity but a decrease of 1.5% in team performance. I will examine this report in more detail separately.

5. WeAreDevelopers “State of WebDev AI” (2025) An interesting survey with 4k+ respondents. Of the web developers surveyed, 91% had experimented with ChatGPT; Cursor IDE led AI editors in awareness, at 80%.

6. The Pragmatic Engineer Survey (2025) An interesting survey with 3k+ respondents. Among engineers, 72% considered AI tooling the “most innovative market”; most noted the rapid arrival of startup tools competing with Copilot.

These studies suggest that AI has become an integral part of software development, while the industry is still adapting and balancing benefits against risks.

P.S. I recently assembled a similar collection (1 and 2) on Developer Productivity for Humans, examining each source individually. I expect to look more closely at the surveys in today’s collection too :)

#AI #ML #Software #Engineering #Processes #Metrics #Management