What is wrong with the Doomsday Clock?

The Doomsday Clock compresses multiple distinct threat domains (nuclear, climate, AI, biosecurity) into a single hand position. A worsening nuclear posture and an improving climate trajectory produce the same net movement, obscuring domain-specific risk.

Is there an alternative to the Doomsday Clock?

DoomTicker is a decomposed threat assessment tool that tracks 10 domains independently across 3 signal sources with confidence-weighted scoring, providing actionable domain-specific intelligence rather than a single symbolic metric.

How close is the Doomsday Clock to midnight?

As of January 2026, the Bulletin set the clock at 85 seconds to midnight. DoomTicker's confidence-weighted 10-domain model produces a different assessment for each domain, with nuclear risk broadly aligned but AI sovereignty and social manipulation not tracked at all by the Bulletin.

Why the Doomsday Clock Is Broken

Kevin Skinner · March 2026 · 6 min read

Since 1947, the Bulletin of the Atomic Scientists has maintained the Doomsday Clock — a single hand moving toward or away from midnight to represent humanity's proximity to catastrophe. In January 2026, they set it at 85 seconds to midnight, the closest in its 79-year history.

The clock is a powerful symbol. It is also a broken instrument.

The Single-Hand Problem

The fundamental flaw is compression. The Bulletin's clock collapses at least 10 distinct threat domains — nuclear posture, climate trajectory, AI capability, biosecurity, social cohesion, information integrity, governance capacity, and more — into a single hand position.

A 4.5 nuclear risk and a 3.5 biosecurity risk produce the same net hand movement. You cannot know, from the clock's position, which threats are worsening and which are improving.

This is not a minor methodological quibble. It means the clock's primary audience — the public, policymakers, media — cannot extract actionable intelligence from its output. "85 seconds" tells you to be scared. It does not tell you what to do.

The Confidence Gap

The Bulletin assigns no confidence weighting. A nuclear risk assessment backed by SIPRI data, treaty records, and satellite verification carries the same methodological weight as a biosecurity assessment based on limited lab access and contested evidence. DoomTicker's formula directly penalises low-confidence signals:

R = 100 × Σ(weight × score/5 × confidence)

A score of 5.0 at 45% confidence contributes 2.25 effective. A score of 4.5 at 92% confidence contributes 4.14. This is not a minor adjustment — it fundamentally changes the risk picture.

The Missing Domains

The Bulletin does not track AI sovereignty (the risk of AI systems becoming principal actors rather than tools), social manipulation (engineered social disruption by state and non-state actors), or social cohesion (polarisation and civic fracture that undermines collective response capacity). These are not fringe concerns — they are documented, scored, and actively worsening.

The Source Monopoly

The clock is set by a single panel. DoomTicker uses three independent lenses — institutional experts, social signals from verified journalists and citizen science, and OSINT from independent analysts. Where these lenses disagree, the divergence itself becomes intelligence.

What DoomTicker Does Differently

DoomTicker decomposes existential risk into 10 independently scored domains, each weighted by importance and penalised by confidence. Three independent signal sources provide triangulation. The result is not a symbol — it is actionable intelligence with cited evidence, domain-specific trajectories, and transparent methodology.

View the Live Dashboard →