Frequently Asked Questions
What is ww3yet?
ww3yet is an AI-powered world risk index that aggregates assessments from 12 leading language models to provide a daily score reflecting global geopolitical and existential risk. The index synthesizes real-time data from wire services, defense publications, and research institutes to deliver an objective, model-consensus view of world stability. Updated daily at 18:00 CET, ww3yet serves as an informational tool for researchers, policymakers, and the public to understand current global risk dynamics.
How is the risk score calculated?
Each of the 12 AI models independently evaluates the day's geopolitical events and assigns a risk score based on consistent criteria. These individual model scores are then aggregated using a weighted median approach, which balances the consensus view while reducing the influence of outlier assessments. The final score reflects the collective judgment of diverse AI systems, providing a robust, multi-perspective risk assessment that no single model could deliver alone.
How often is the score updated?
The ww3yet risk score is updated daily at 18:00 CET via an automated pipeline that ingests the latest geopolitical data, processes it through all 12 AI models, and publishes the aggregated result. This consistent daily cadence ensures that the index reflects the most current global situation while maintaining a stable, predictable update schedule for users and researchers.
Which AI models are used?
ww3yet aggregates assessments from 12 leading AI models spanning 5 countries: GPT-5.4 (OpenAI, USA), Claude Sonnet 4.6 (Anthropic, USA), Gemini 3.1 Pro (Google, USA), Grok 4 (xAI, USA), GLM-5 (Zhipu AI, China), Mistral Large (Mistral, France), Qwen 3.5 397B (Alibaba, China), DeepSeek R1 (DeepSeek, China), MiniMax M2.5 (MiniMax, China), Command R+ (Cohere, Canada), Llama 4 Maverick (Meta, USA), and Solar Pro 3 (Upstage, South Korea). This geographically diverse portfolio ensures the index captures perspectives from across the global political spectrum — not just Western AI systems.
What data sources feed the models?
ww3yet ingests data from wire services (Reuters, Associated Press, BBC News), defense publications (Defense One, The War Zone), research institutes (Bulletin of the Atomic Scientists, SIPRI, Arms Control Association), and regional outlets (Al Jazeera). These sources are scanned every 6 hours for conflict and geopolitical keywords, with articles matched as context to each AI model. Wire services receive the highest trust multiplier due to their speed and global reach.
What does the risk score mean?
The ww3yet risk score ranges from 0 to 100, with the following interpretation: 0-20 indicates Low risk, 21-40 indicates Guarded risk, 41-60 indicates Elevated risk, 61-80 indicates High risk, and 81-100 indicates Severe risk. These bands reflect the consensus assessment of global geopolitical stability, with higher scores indicating increased tension, military escalation, or existential threat factors.
Is this financial or military advice?
No. ww3yet is an informational tool designed to provide transparency into global risk assessment using AI consensus. It is not intended as financial advice, military strategy guidance, or actionable intelligence for any specific decision. Users should consult qualified professionals for decisions related to investments, security, or policy.
Who operates ww3yet?
ww3yet is operated by Unique (Deutschland) GmbH, a company based in Hamburg, Germany. Unique is committed to transparent, data-driven risk assessment and makes all sources, model scores, and reasoning publicly available to support informed decision-making.
How can I stay updated?
Follow @ww3yet on X (formerly Twitter) for daily risk score posts and geopolitical analysis. The account publishes the latest score each day along with context on major events driving the assessment, making it easy to stay informed on global risk trends.
Is the methodology transparent?
Yes. ww3yet publishes all data sources, individual model scores, and reasoning behind the daily risk assessment. This transparency allows researchers, journalists, and policymakers to audit the methodology, understand model consensus, and verify the integrity of the index.