Reported artificial intelligence incidents rose 55 percent in a single year, climbing to 362 documented incidents in 2025 from 233 in 2024, according to the Stanford AI Index and the AI Incident Database it draws on. The increase tracks with rapid AI adoption and a widening gap between deployment growth and the governance and safety measures meant to keep pace.

A separate monitoring effort recorded an even higher tempo. The OECD's AI Incidents and Hazards Monitor, which uses a broader automated collection pipeline, recorded a peak of 435 incidents in a single month in January 2026, with a six-month moving average of 326. The two data sources differ in methodology, but both point to a sharp rise in documented AI failures and controversies.

The distribution of incidents also shifted. Among organizations experiencing incidents, the share reporting three to five incidents rose from 30 percent to 50 percent, while the share with only one or two fell from 42 percent to 29 percent, indicating that incidents are becoming recurrent within the same organizations, often those adopting AI most aggressively. The Stanford report, published in April 2026, frames the trend as evidence that governance and safety practices are lagging behind the pace of AI capability and deployment.

Source: Stanford HAI -- https://hai.stanford.edu/ai-index/2026-ai-index-report/responsible-ai