The AI Incident Database recorded 362 documented AI incidents in 2025, up from 233 in 2024, according to the 2026 AI Index Report published by the Stanford Institute for Human-Centered AI. The jump amounts to a 55 percent year over year increase.

The report sets that count against uneven measurement of AI system behavior. Almost all leading frontier model developers publish results on capability benchmarks such as MMLU and SWE-bench, while reporting on responsible AI benchmarks remains sparse. On a new accuracy benchmark cited in the report, hallucination rates across 26 top models ranged from 22 percent to 94 percent.

Disclosure scores moved in the opposite direction from incident counts. The average score on the Foundation Model Transparency Index rose from 37 to 58 between 2023 and 2024, then dropped to 40 in 2025. The report identifies persistent gaps in disclosure around training data, compute resources, and post-deployment impact.

Corporate governance structures expanded over the same period. AI-specific governance roles grew 17 percent in 2025, and the share of businesses operating with no responsible AI policies fell from 24 percent to 11 percent. Organizations named knowledge gaps at 59 percent, budget constraints at 48 percent, and regulatory uncertainty at 41 percent as the leading obstacles to putting those policies into practice.

Incident totals in the database reflect publicly reported events, so year over year movement tracks reporting behavior alongside the underlying rate of system failure.

Source: Stanford Institute for Human-Centered AI - https://hai.stanford.edu/ai-index/2026-ai-index-report/responsible-ai