Documented artificial intelligence incidents rose to 362 in 2025, up from 233 in 2024, according to the AI Incident Database, a sharp year-over-year increase in reported failures, harms, and controversies. The database had logged roughly 800 to 900 unique incidents through the first quarter of 2026, growing about 35% to 45% year over year, faster than AI deployment itself.

Generative AI has become the dominant source of new entries. Systems that produce text, images, and video accounted for about 58% of 2025 incidents, reversing the earlier prevalence of recommendation-system and computer-vision cases. The shift tracks the rapid spread of generative tools across business and consumer use.

The largest incident categories center on content and fairness. Misinformation and content harms make up about 28% of logged incidents, discrimination and bias about 22%, and physical safety failures about 14%. Separately, the OECD AI Incidents and Hazards Monitor reached 435 reports in January 2026, with a six-month moving average of 326.

Analysts note that rising incident counts partly reflect improved detection and reporting rather than deployment alone. The data indicates that as organizations adopt AI more widely, the documented record of failures is expanding quickly, giving companies and regulators a growing evidence base on where these systems break down.

Source: Statista -- https://www.statista.com/chart/35846/ai-incidents-involving-content-generation/