• Questions
  • Tournaments
  • Services
  • News
  • Questions
  • Tournaments
  • Questions
  • Questions
14 comments
85 forecasters

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

Google DeepMind13%
Microsoft5%
xAI3%
No key factors yetAdd some that might influence this forecast.
Add key factor
Authors:
will_aldred
Opened:Mar 12, 2024
Closes:Jan 1, 2026
Scheduled resolution:Jan 1, 2026
AI Safety
Future of AI
Computing and Math
Artificial Intelligence
2024-2025 Leaderboard
Psychology Today logo
Double Tipping Points: Agency Decay in the Climate-AI Nexus
Psychology Today•Sep 15, 2025
TechCrunch logo
‘Selling coffee beans to Starbucks’ – how the AI boom could leave AI’s biggest companies behind
TechCrunch•Sep 14, 2025
Business Insider logo
Anthropic bot crawlers feast on web content and give little back, a new ranking shows
Business Insider•Sep 12, 2025
Learn more about Metaculus NewsMatch

Will there be a frontier open-source AI model on January 1 of the following years?

2024
No
2025
Yes
2026
63.1%
1 other
125 forecasters

Will any frontier model be released in 2029 without a third-party evaluation of dangerous capabilities?

59% chance
48

Will there be a leading AI lab with no internal safety team in the following years?

2026
64%
2030
37%
2036
20%
35 forecasters
  • About
  • API
  • FAQ
  • forecasting resources
  • For Journalists
  • Careers
GuidelinesPrivacy PolicyTerms of Use
ForbesScientific AmericanTimeVoxYale NewsNature