08 / SOURCES
Where the data comes from
Adoption estimates differ by 10–40+ points depending on question wording and sampling. This table shows what each survey actually measures — so you can judge comparability yourself.
| Source | Sample | Key Metric | What It Measures |
|---|---|---|---|
| Stack Overflow 2023 | 89,184 | 43.8% using AI tools | "Use in development process" — narrower than "ever used" |
| Stack Overflow 2024 | 65,437 | 61.8% currently using | Same framing as 2023; year-over-year comparable |
| Stack Overflow 2025 | 49,009 | 47.1% daily, 78.5% total | Shifted to frequency reporting; trust data added |
| DORA 2025 (Google) | ~5,000 | 90% adoption, 2 hr/day median | Technology professionals (not dev-only); org-level outcomes |
| HackerRank 2025 | 13,732 | 97% use ≥1 AI assistant | "At least one" = low bar; hiring community sample |
| JetBrains 2025 | 24,534 | 85% regularly use AI tools | IDE ecosystem survey; "regularly" undefined |
| GitHub Survey 2024 | 2,000 | ~97% ever used AI tools | "Ever used at any point" — not a frequency measure |
| Gartner 2024 | 598 | 63% orgs piloting/deploying | Enterprise IT lens; forecasts, not direct adoption |
| Pragmatic Engineer 2026 | 906 | Claude Code 71% among agent users | Small, self-selected; median 11–15 yrs experience |
Survey Comparison Matrix
Side-by-side view highlighting methodological differences that explain why adoption numbers range from 47% to 97%.
| Dimension | Stack Overflow 2025 | DORA 2025 | HackerRank 2025 | GitHub 2024 |
|---|---|---|---|---|
| Headline adoption | 78.5% | 90% | 97% | ~97% |
| Question framing | "Currently using" (frequency) | "Use AI tools" (any) | "Use at least one AI assistant" | "Ever used AI coding tools" |
| Sample size | 49,009 | ~5,000 | 13,732 | 2,000 |
| Audience | Self-selected developers | Tech professionals (broader) | Hiring community | Commissioned panel |
| Selection bias | SO users skew experienced | Google-adjacent orgs | Job-seekers, early career | Paid participants |
| Trust data | Yes (46% distrust) | Yes (24% "a lot") | No | No |