Holden Karnofsky
Co-founder and former co-CEO
Open Philanthropy
Safety researchers
Co-architect of the most-important-century case: that we may live in the hinge of history because AI is likely to transform everything within our lifetimes.
Their number
10–30%
I put the chance of an AI-related catastrophe this century at more than 1 in 10.Source
On timing
Argues transformative AI this century is more likely than not, and possibly within the next decade or two.
SourceWhat they call for
- Massive expansion of alignment research
- Standards and evaluation regimes for frontier models
- Funder and lab norms for caution