No. 458 of 500 No. 54 in AI Safety & the Future

Global Catastrophic Risks

By Nick Bostrom and Milan M. Ćirković

Published
August 1, 2011
Print length
576 pages
Reading level
Advanced
ISBN-13
978-0199606504
Publisher
Oxford University Press
Written for
Researchers, policy analysts and students who want the intellectual foundations of existential-risk studies.

Synopsis

Nick Bostrom and Milan Ćirković assembled 25 authors across 22 chapters to treat risks that could kill around ten million people or cause ten trillion dollars in damage as a single field of study. Opening chapters set the scale, from the long-term fate of the universe to human evolution, then examine how judgment goes wrong: end-of-the-world beliefs, cognitive biases in risk assessment, and the observation selection effect, which can lead people to underestimate extinction-level events because our own existence implies that none has struck in our past. Further chapters offer tools for response, including systems risk management, insurance methods and Richard Posner's cost-benefit case for spending far more on prevention.

The remaining chapters take thirteen risks in turn, grouped as risks from nature (supervolcanoes, asteroid and comet impacts, supernovae and gamma-ray bursts), from unintended consequences (climate change, pandemics, artificial intelligence, physics experiments, social collapse) and from hostile acts (nuclear war, nuclear terrorism, biotechnology, nanotechnology, totalitarianism). Eliezer Yudkowsky's chapter on artificial intelligence argues that premature confidence in understanding AI is itself the main hazard and makes the case for deliberately Friendly design. Readers gain a shared vocabulary for comparing very different catastrophes and a record of how the existential-risk field defined itself before deep learning.

Scorecard

Editorial score

How it scored

Editorial score: 66.3 / 1005 weighted criteria

  • Influence30%

    22.2/30
  • Authority25%

    21.0/25
  • Practical value20%

    10.4/20
  • Currency15%

    7.2/15
  • Readability10%

    5.5/10
  • Editorial score

    66.3/100

How the rank was computed

Rank No. 458 of 500 comes from a composite of 48.3: observable signals 18.4 of 55 points plus editorial review 29.9 of 45 points.

All 11 inputs
InputPoints
Observable signals18.4/55
Reach (Amazon ratings volume)6.8/16.5
Consensus across best-of lists5.5/13.8
Awards and bestseller evidence0.0/8.3
Reader rating, sample-adjusted5.6/8.3
Amazon sales rank0.5/5.5
Recency0.0/2.8
Editorial review29.9/45
Author authority9.5/11.3
Influence10.0/13.5
Practical value4.7/9.0
Readability2.5/4.5
Currency3.2/6.8

Each criterion is scored from 1 to 10 and weighted, and the points add up to the editorial score. The rank combines observable signals (55 percent) with this score (45 percent). Read the methodology.

Verdict

The first major edited survey of civilization-scale threats, notable for an early AI-risk chapter, now read mainly for its framing and history.

Strengths

  • Broad coverage of thirteen distinct catastrophic risks
  • Annotated further-reading lists for every chapter
  • Contributors span physics, law, economics and psychology

Keep in mind

Chapter quality is uneven, and the AI, biotechnology and climate chapters reflect 2008 knowledge, so readers need newer work for current assessments.

Key takeaways

  1. Studying catastrophic risks together reveals shared patterns, such as dual-use technologies, and helps show which threats deserve priority attention and funding.
  2. Observation selection effects can make humanity underestimate extinction-level risks, because our existence implies none has yet occurred in our own past.
  3. Yudkowsky argues that the greatest AI danger is premature confidence that we understand the technology, compounded by cognitive biases and scarce data.
  4. Posner's cost-benefit chapter contends that even modest valuations of human survival justify far larger investment in catastrophe prevention than societies now make.

Notable facts

  • Nature reviewed the book in October 2008 in a piece titled 'Burning down the house' by Bill McGuire, director of the Benfield UCL Hazard Research Centre. Source (opens in a new tab)
  • Eliezer Yudkowsky's chapter 'Artificial intelligence as a positive and negative factor in global risk' occupies pages 308–345 of the volume. Source (opens in a new tab)
  • Seth Baum's review for Risk Analysis called the book an invaluable resource for anyone working on global catastrophic risk. Source (opens in a new tab)

About the author

Nick Bostrom and Milan M. Ćirković

Nick Bostrom is a Swedish philosopher who founded Oxford University's Future of Humanity Institute in 2005 and led it until its closure in 2024; he wrote Superintelligence (2014) and Deep Utopia (2024). Co-editor Milan M. Ćirković is a research professor at the Astronomical Observatory of Belgrade and author of The Great Silence, a study of the Fermi paradox.

Questions

What is Global Catastrophic Risks about?
It is an edited academic volume, led by Nick Bostrom and Milan Ćirković, that surveys threats capable of causing death and damage on a global scale. Chapters cover natural hazards such as asteroid impacts and supervolcanoes, unintended dangers such as pandemics and artificial intelligence, and hostile acts such as nuclear war and bioweapons, along with methods for assessing and managing such risks.
Does Global Catastrophic Risks cover artificial intelligence?
Yes. Eliezer Yudkowsky's chapter, Artificial Intelligence as a Positive and Negative Factor in Global Risk, argues that the greatest danger is concluding too early that we understand AI, and it explains why deliberately safe design matters. Written in 2008, well before deep learning took hold, it is an early academic statement of arguments that later shaped AI alignment research.
Is Global Catastrophic Risks worth reading in 2026?
It is worth reading for anyone who wants the origins of existential-risk studies and a single framework for comparing very different threats. Many chapters, particularly on AI, biotechnology and climate, reflect knowledge from 2008, so the book works best as a foundation paired with recent research rather than as a current risk assessment.

Sources

As an Amazon Associate, Best AI Books earns from qualifying purchases. Best AI Books is sponsored by Iternal.ai, whose founder and CEO, John Byron Hanby IV, wrote The AI Strategy Blueprint, our #1 ranked book. How we rank books.