17
66 forecasters

Will leading AI labs have their models evaluated for dangerous behavior before 2026?

Amazon
result: Yes
Google DeepMind
Annulled
Microsoft
result: No
Meta
result: No