88% of organizations worldwide used AI in 2025, up from 55% in 2023, according to The 2026 AI Index Report published by the Stanford Institute for Human-Centered Artificial Intelligence (Stanford HAI) in April 2026. The report tracks what happened in AI during 2025.
About the report
| Publisher | Stanford Institute for Human-Centered Artificial Intelligence (Stanford HAI) |
| Released | April 2026 |
| Title | The 2026 AI Index Report |
| Data covered | Research, technical performance, responsible AI, economy, education, policy and public opinion, mainly for 2025 |
| Latest edition? | Yes, as of November 2026 (annual) |
Key findings
- Global corporate investment in AI reached a record $581.69 billion in 2025, including $344.66 billion in private investment, according to Stanford HAI.
- The share of organizations using AI rose to 88% in 2025 from 55% in 2023, with Europe leading regions at 91%, the AI Index reports.
- Generative AI reached 53% adoption within three years, faster than the personal computer or the Internet at the same stage, according to Stanford HAI.
- On OSWorld, a benchmark of real computer tasks, AI accuracy rose to 66.3% in 2025, still below the human baseline of 72.35%, the report finds.
- Reported AI incidents hit a record 362 in 2025, according to the AI Index.
- 32% of organizations expect AI to reduce their workforce next year, against 13% expecting an increase and 43% expecting little or no change, Stanford HAI reports. Headcounts for software developers and customer support agents aged 22 to 25 fell between 2021 and 2025.
| Indicator (Stanford HAI) | Value |
|---|---|
| Organizations using AI, 2023 | 55% |
| Organizations using AI, 2025 | 88% |
| Generative AI adoption after three years | 53% |
| AI accuracy on OSWorld, 2025 | 66.3% |
| Human baseline on OSWorld | 72.35% |
What this means for AI and data learners
Using AI at work is now the default. When 88% of organizations use it, fluency with AI tools belongs on a CV the way spreadsheet skills did twenty years ago. Pick one task in your daily work, run it with an AI tool, and measure the time saved.
Entry-level tech roles are under pressure. Falling headcounts among developers and support agents aged 22 to 25 send a clear signal: routine tasks are shrinking, and value moves to people who combine technical skills with the judgement to review and steer AI output.
Agents are closing in on human performance. A jump to 66.3% accuracy on real computer tasks means agent-based automation is becoming practical for small teams. Follow how these platforms are evolving in our piece on ChatGPT becoming an app platform.
Safety is part of the skill set. With 362 reported incidents in one year, understanding risks such as data leaks and wrong outputs is part of professional competence, not an optional extra.
📥 Full report: Read the 2026 AI Index Report for free on the Stanford HAI website.
FAQ
The Stanford Institute for Human-Centered Artificial Intelligence (Stanford HAI), every year since 2017.
Mainly developments in AI during 2025.
$581.69 billion in global corporate AI investment, including $344.66 billion in private investment, according to the report.
Yes, the 2026 edition is the latest as of November 2026.
