| Canonical name | MIT AI Incident Tracker |
| Address | https://airisk.mit.edu/ai-incident-tracker |
| Operator | MIT AI Risk Initiative, with Arcola AI's LLM pipeline op |
| Kind | database |
| Roles | harm |
| Lifecycle | active |
| Observation | 2026-10-28: read |
| Unit | AI Incident Database incident, classified by risk domain, cause, harm, severity and EU AI Act risk level op |
| Intake | AIID reports processed by an LLM pipeline (Arcola AI) against the MIT Risk Repository taxonomy and a severity rating based on CSET's AI Harm Taxonomy op |
| Count | 1,600 (the tracker page's heading) and 'more than 1,400' (its description); Google's card summary of 2026-10-29 gives 1,752; the MIT AI Risk Navigator's heading gives 1,497 (2026-10-29) (as of 2026-10-29) obs |
| Note | the page states that patterns 'should be taken as indicative'; reporting is voluntary and no systematic validation study is complete op |
op operator statement · 3p third-party report · obs our observation · rd our reading
It gives AIID's records a more consistent structure for policymakers, by its own account. The classification is machine-made: an LLM pipeline reads the raw AIID reports and assigns the MIT taxonomy and a severity rating based on CSET's AI Harm Taxonomy.
The page gives its size twice, as "more than 1,400" in its description and 1,600 in its heading; the MIT AI Risk Navigator gives 1,497. It notes that reporting is voluntary, that no systematic validation study is complete, and that patterns "should be taken as indicative."
It is a derived layer: the editor-resolved incidents of AIID, reclassified by a model. The chain from event to news report to AIID record to machine classification is a chain of receptions, each recorded.
| Admission requires harm? | inherits AIID's harm test rd |
| Preserves original outputs? | no: classifications are machine-written rd |
Lee Sharks, ed. "MIT AI Incident Tracker." MMRS meta-aggregator, Machine-Mediated Reception Studies, compiled 2026-10-29. https://www.machinemediation.org/aggregators/mit-ai-incident-tracker/
Record https://www.machinemediation.org/aggregators/mit-ai-incident-tracker/ · entity https://www.machinemediation.org/aggregators/mit-ai-incident-tracker/#entity