⚠️
⚠️
RESULTS PREVIEW
Please do not share except with trusted colleagues.
Email us at
airisk@mit.edu
with any questions or feedback.
AI Risk Actor Vulnerability Matrix
Round 3 Data
All Risks Visible
Select All
Deselect All
All Actors Visible
Select All
Deselect All
Show Non-Consensus Responses
Show Low-Evidence Responses (n < 10)
Vulnerability Levels
Not at all vulnerable
Minimally vulnerable
Moderately vulnerable
Highly vulnerable
Extremely vulnerable
Consensus Indicator
✓
Checkmark (✓) indicates expert consensus reached
Consensus definition: ≥90% of responses within ±1 of median AND ≥60% exactly on median
Vulnerability Assessment
×
×
×