The p(doom) database · person
David Duvenaud’s p(doom)
~85%
In their words
David Duvenaud (Professor, University of Toronto; former Anthropic researcher) put their p(doom) at ~85% (p(doom), as asked on Doom Debates), on a podcast on Jul 29, 2026.
So I'd say something around 85%.
Source · re-checked against the primary source.
Clip replayed in the State of the Show episode; originally from his own Doom Debates episode.
Every number on the record
| Date | p(doom) | Scope | Source |
|---|---|---|---|
| Jul 29, 2026 | ~85% | p(doom), as asked on Doom Debates | podcast |
| Aug 21, 2025 | 70–80% | 'doom' as a state where almost everything he values has been destroyed, even if humans survive (gradual disempowerment) · by 2100 | podcast |
In context
Among ex-frontier labs in the database (8 people), the median is 12.5%. Across all 223 people it is 10%. How the medians and the index are built.
A p(doom) is a personal, subjective estimate of the probability that advanced AI ends in human extinction or a comparable, irreversible catastrophe. Scope and timeframe differ from person to person, so compare with care: why p(doom) numbers disagree.