Alex Turner’s p(doom)
In their words
Alex Turner (AI safety researcher, formerly Google DeepMind) put their p(doom) at ~1 in 3 (AI takeover of human civilization), in an article on Sep 14, 2026.
I myself would guess AI takeover chances at roughly one-in-three – not a coin flip, but high enough to justify urgent action.
Source · re-checked against the primary source.
Guardian op-ed. His Aug 2026 Doom Debates figure (25–30%) was for AI killing at least a billion people by 2050; this is takeover. Restated at the NYC Council hearing, Oct 5, 2026 ("roughly one in three", as reported by Fortune and The Next Web). On frominside.ai (Sept 29), conditional: "Assuming an AI has taken control, it seems … over fifty percent more likely than not that all humans would end up dying."
Every number on the record
| Date | p(doom) | Scope | Source |
|---|---|---|---|
| Sep 14, 2026 | ~1 in 3 | AI takeover of human civilization | article |
| Aug 4, 2026 | 25–30% | AI kills at least a billion people · by 2050 | podcast |
In context
Among ex-frontier labs in the database (9 people), the median is 27.5%. 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.