------
PKT 000000
NODES 15
SIG
98.4%
APX UPLINK
ENCRYPTED SESSION: ACTIVE
ALPHA KEY --------
UTC --:--:--

FREQUENTLY ASKED QUESTIONS

01 What is the Nuclear War Score?
The Nuclear War Score is a daily aggregated assessment from 0 to 20 that reflects the perceived global nuclear threat level as evaluated by leading AI models.
02 How is the score calculated?
Each day, major AI models are independently queried. Their responses are averaged to produce a single composite score for the day.
03 What does a score of 0 mean?
A score of 0 indicates no perceived nuclear threat according to the AI consensus. This is the lowest possible rating on the scale.
04 What does a score of 20 mean?
A score of 20 represents the maximum perceived nuclear war risk. This would reflect extraordinary global tensions as assessed by all queried AI models simultaneously.
05 Is this score an official government assessment?
No. This is an independent, experimental project aggregating AI opinions. It should not be interpreted as an official, scientific, or governmental evaluation.
06 How often is the score updated?
The score is updated once per day. Historical scores are preserved and displayed on the [LOGS] page.
07 Which AI models contribute to the score?
Several models are queried: ChatGPT, Claude, Gemini, Llama, DeepSeek, Mistral, Qwen, Gemma, Nvidia, Arcee, Grok, Perplexity, Cohere, Jamba and Moonshot.
08 How are my personal data stored?
Visit our [PRIVACY POLICY] page to know more.
09 What is the score level classification?
Scores are grouped into five levels: LOW (0–4), GUARDED (5–9), ELEVATED (10–13), HIGH (14–17), and CRITICAL (18–20). These are illustrative labels, not official categories.
10 What is the AI Disagreement Index?
The disagreement index is the standard deviation of all model scores. A low value means models strongly agree; a high value means models diverged significantly in their assessment.
11 Can I use or share this data?
Yes. The data is freely accessible. We ask that you credit this site when sharing and remind your audience that it represents AI-aggregated opinions, not scientific fact.You may also query existing data via our [PUBLIC API], and subscribe to our [RSS FEED].
12 How are the dashboard metrics calculated?
The dashboard combines multiple AI model evaluations into unified intelligence signals. Each metric represents a different view of risk, agreement, and movement across categories.

Key metrics explained:
  • Global Signal — Overall narrative based on which categories rank highest and lowest in risk.
  • Global Score Index — Average risk across all categories (sum of scores ÷ number of categories).
  • Global Score Change (24h) — Difference between today’s and yesterday’s global risk index.
  • Score Ranking — Categories ordered from highest to lowest average risk score.
  • Dominance Gap (Top vs Runner-up) — Difference between the highest and second-highest category scores.
  • Momentum Shift — Per-category change compared to yesterday’s score.
  • Biggest Mover — Category with the largest absolute change in risk level.
  • Chaos Index — Average disagreement between AI models across all categories.
  • Sentiment Bias — Overall tone of model outputs based on aggregated scoring (optimistic, neutral, pessimistic).
  • Consensus Level — How aligned models are within each category based on disagreement scores.
  • Model Outliers — Models that significantly deviate from the category average.
  • Top Signal Drivers — Most frequent reasons influencing model scores, aggregated across all categories.

All values are derived from AI-generated assessments and are meant for analytical visualization, not scientific prediction.