Nick Bostrom
Philosopher; author of 'Superintelligence' (2014) and 'Deep Utopia' (2024); founder of the former Future of Humanity Institute
Macrostrategy Research Initiative (formerly University of Oxford FHI)
Safety researchers
Bostrom's 2014 book 'Superintelligence' put AI extinction risk on the map for Silicon Valley. A decade on he still takes the danger seriously but now warns against over-caution too, arguing that never building superintelligence would itself be a catastrophe and that humanity should 'feel our way through' the transition.
Declines to give a number
No current numeric p(doom). Recognises 'significant existential risks' but has recently stressed the opposite failure: "It would be in itself an existential catastrophe if we forever failed to develop superintelligence" (March 25, 2025).Source
On timing
Takes short timelines seriously (single-digit years to superintelligence) while allowing for 10-20 years or longer (2025 interviews).
SourceWhat they call for
- Proceed carefully to superintelligence rather than abandon it
- Possible pauses to install safeguards during competitive races
- 'broadly cooperative future' respecting many values
- Attention to AI welfare/moral status
In their own words
"I think ultimately this transition to the superintelligence era is one we should do." / "We'll have to feel our way through this and make adjustments as we go along."
25 March 2025
Governance might require "pauses in development to install safeguards amid competitive races" (Deep Utopia-era interviews).
24 August 2025
Signed
- CAIS extinction statement 2023 (reported; not independently re-verified on this pass)