P(doom) Calculator

What's your personal probability estimate of catastrophic AI outcomes? Calibrate your existential risk view with structured thinking.

P(doom) calculator — AI existential risk probability estimator

⚠️ This is an educational and exploratory tool. These estimates are deeply uncertain. Experts range from <1% to >50% P(doom). Use this to structure your thinking, not as a definitive answer.

Risk Factor Assessment

70%
30%
40%
10%
30%
Your P(doom)
0%
Alignment path
0%
Malicious use path
0%

Expert Estimates (for reference)

Eliezer Yudkowsky (MIRI): ~99%
Geoffrey Hinton: ~10–20%
Paul Christiano (ARC): ~10–20%
Yann LeCun (Meta): <1%
Sam Altman (OpenAI): Low, manageable
2022 AI Researcher Survey median: ~5%
P(doom) = 1 − (1 − Path₁) × (1 − Path₂)
Path₁ = P(AGI) × P(Misalignment | AGI) × P(Catastrophe | Misalignment) × (1 − P(Governance)) · Path₂ = P(AGI) × P(Malicious use) × (1 − P(Governance))
>99%

Yudkowsky

MIRI founder — most pessimistic major researcher

10–20%

Hinton / Christiano

Geoffrey Hinton & Paul Christiano (ARC) estimates

~5%

Median researcher

2022 AI researcher survey median P(doom)

<1%

LeCun

Yann LeCun (Meta) — most optimistic major researcher

Uncertainty Quantification

The 100× Disagreement Matters

When experts disagree by two orders of magnitude (<1% vs >99%), the takeaway is not that one side is right — it's that we face genuine uncertainty about unsolved problems. P(doom) is a tool for structuring uncertainty, not a prediction. The most rational response is to support AI safety research while acknowledging the limits of our current knowledge.

Why Experts Disagree by 100×

P(doom) estimates range from <1% to >99% because they depend on assumptions about unsolved open problems: whether the alignment problem is tractable, how quickly AI capabilities will advance, and whether governance mechanisms can keep pace with technology.

These are not empirical disagreements resolvable by more data — they are partly philosophical judgements about the nature of intelligence, human institutions, and the behaviour of optimisation processes at scale.

Educational tool: This calculator structures your uncertainty — it does not produce a ground truth. P(doom) is not a scientific metric with a knowable true value. Use it to explore your own assumptions.

What Drives Estimates Up or Down

Factors that increase P(doom)

Rapid capability scaling outpacing safety, race dynamics between labs or nations, difficulty of the alignment problem, inadequate governance, and dual-use misuse potential of powerful models.

Factors that decrease P(doom)

Capability plateau before dangerous levels, alignment proving tractable, international cooperation, significant lab safety investment, democratic oversight, and narrow AI remaining dominant longer than expected.

What you can do

Engage with AI governance debates. Support responsible AI institutions (ARC, MIRI, Anthropic safety teams, Centre for AI Safety). Stay informed. The civic response is more impactful than individual behaviour changes.

Frequently Asked Questions

P(doom) is the estimated probability that advanced AI development leads to catastrophic or existential outcomes for humanity. It is a heuristic used by AI safety researchers, not a formal scientific metric. Estimates range from <1% to >99%, reflecting genuine and deep disagreement about unsolved problems.
Geoffrey Hinton estimates 10–20% chance of existential risk. Eliezer Yudkowsky (MIRI) expresses P(doom) > 99%. Paul Christiano (ARC) estimates ~10–20%. Stuart Russell regards the risk as serious and motivating. Yann LeCun (Meta) considers current AI architectures incapable of posing such risks. The disagreement is real and substantive.
Rapid capability advancement leaving safety work behind, lack of alignment progress, competitive race dynamics between nations and companies, the difficulty of the alignment problem, inadequate governance, and misuse potential by malicious actors with access to powerful models.
AI capability plateaus before dangerous levels, alignment proving tractable, international cooperation developing, significant AI lab safety investment, democratic oversight mechanisms, and narrow AI remaining dominant far longer than pessimists expect.
This calculator is a thought experiment and educational tool, not actionable risk management. P(doom) estimates are a prompt to engage with AI governance debates and support safety research. Most impactful actions: engage with AI policy, support responsible AI institutions, stay informed.
The alignment problem is the challenge of ensuring that advanced AI systems pursue goals that are beneficial to humanity — even as they become more capable. It is hard because we cannot perfectly specify human values, and because capable optimisers may find unexpected ways to satisfy specified goals that violate the spirit of our intentions.
Artificial General Intelligence (AGI) is typically defined as AI that can perform any cognitive task a human can. Estimates of arrival range from 2028 (optimistic) to never (sceptics). The 2022 AI Impacts survey found median researcher estimates of ~50% chance by 2059. Timelines have shortened significantly since GPT-4.
Yes, meaningfully. Constitutional AI (Anthropic), RLHF fine-tuning, interpretability research (Mechanistic Interpretability), and scalable oversight are active areas. Whether progress is fast enough relative to capability scaling is the core dispute. Funding for safety research has grown substantially since 2022.

For informational purposes only — not financial, medical, or legal advice. Results are estimates; use at your own risk. Full terms