AI Incident Tracker

This dashboard shows the distribution of risks within each year, classified by the MIT Causal Taxonomy, the MIT Domain Taxonomy and the EU AI Act Risk Levels.

Incidents prior to 2015 not shown.

Example use-case question:

• Which risk domains are trending towards increased proportions of reported incidents?
• Is the proportion of intentionally caused AI safety incidents increasing or decreasing as a trend?
• Is the proportion of incidents where the risk would be classed as 'Unacceptable' under the EU AI Act increasing?

‍