Eli Lifland’s p(doom)
In their words
Eli Lifland (AI Futures Project; co-author, AI 2027) put their p(doom) at ~67% takeover (if no slowdown) (AI takeover, if AI companies don't slow down; human extinction ~40% conditional on takeover), on X on Oct 11, 2026.
If they don't slow down, AI takeover is ~67%. Conditional on AI takeover, human extinction is ~40%. Many non-extinction worlds still have lots of deaths.
Source · re-checked against the primary source.
Takeover up from ~50% (Apr 2025). Multiplied out, about 27% extinction if there is no slowdown (our arithmetic, not his). Same post: superhuman AI ~10–15% by end of 2027, ~50% by end of 2030.
Every number on the record
| Date | p(doom) | Scope | Source |
|---|---|---|---|
| Oct 11, 2026 | ~67% takeover (if no slowdown) | AI takeover, if AI companies don't slow down; human extinction ~40% conditional on takeover | x |
| Apr 10, 2025 | ~25% | extinction; misaligned AI takeover ~50% | article |
| Feb 1, 2023 | 35–40% | existential catastrophe from AI · by 2070 | forum |
In context
Among safety orgs in the database (33 people), the median is 50%. Across all 252 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.