◆ Threat History

Showing up to 100 most recent assessment cycles.

◆ How It Works

The AI Dominance Monitor tracks how far AI has penetrated the world across four dimensions: Labor, Military, Government, and Media. It runs a full assessment cycle daily at 10 AM EST.

Step 1 — Live Signal Collection

Each cycle fetches real-time data from three public APIs in parallel:

• GitHub — count of AI/ML repositories with active pushes in the last 30 days.
• HuggingFace — new text-generation model releases in the last 30 days.
• Wikipedia — monthly page views on the Artificial Intelligence and AGI articles.

Step 2 — Web Searches

Claude Sonnet runs 5 web searches to find what is actually happening in AI right now:

• General AI news — open search to surface whatever is dominating coverage this cycle.
• Jobs — AI automation layoffs and workforce displacement.
• Military — autonomous AI weapons deployments and combat use.
• Government — AI surveillance, regulation, and policy developments.
• Media — deepfakes, disinformation, and AI influence operations.

Step 3 — Evidence Classification

Every search result is classified as a Confirmed Event, Data Point, Projection, or Commentary. Only Confirmed Events and Data Points move scores. Commentary is ignored entirely — a headline saying “AI could displace millions” has zero effect on the jobs score.

Step 4 — Calibrated Scoring

Scores are anchored to defined real-world thresholds. A score of 50 means AI is significant but humans retain dominant control. A score of 75 means AI is operating at scale with frequent human bypass. A score of 90 means near-total AI dominance. Scores only move when confirmed evidence warrants it — upward drift without evidence is treated as a bias failure.

Probability Gauges

Each gauge answers a specific measurable question. Scores are grounded in current evidence and expert consensus but are not capped — if confirmed developments warrant higher estimates, they will reflect that.

• AGI Arrival — probability of a publicly deployed AI system performing at or above median human level across the majority of economically valuable cognitive tasks before 2030. Current expert consensus anchors around 10–40%, used as a starting point not a ceiling.
• Existential Risk — probability that AI causes death or permanent incapacitation of more than 1 billion people before 2100. Stays low absent a confirmed specific catastrophic pathway, but is not artificially suppressed.
• Mass Unemployment — probability that any G7 economy sees net AI-driven job loss exceeding 10% of its workforce within 10 years. Anchored to current BLS data and confirmed displacement research.

Intelligence Feed

Headlines are credibly reported events from named outlets only — not opinion or speculation.

This tool is for informational and educational purposes only. All scores are AI-generated estimates, not verified intelligence or predictions.

◆ Terms of Use

Last updated: April 2026

1. Informational Purpose

The AI Dominance Monitor is provided for informational and educational purposes only. All scores, probability estimates, threat assessments, and narratives are generated by artificial intelligence and should not be treated as factual predictions, investment advice, or actionable intelligence.

2. No Warranties

This tool is provided “as is” without warranties of any kind. We do not guarantee the accuracy, completeness, reliability, or timeliness of any data displayed. AI-generated content may contain errors, hallucinations, or outdated information.

3. Data Sources

The Monitor aggregates data from third-party public APIs (GitHub, HuggingFace, Wikimedia) and AI-powered web searches via Anthropic’s API. We are not responsible for the accuracy or availability of these external sources. This site is hosted on Cloudflare Pages; by using this site you are also subject to Cloudflare’s Terms of Service.

4. Limitation of Liability

In no event shall the creators be liable for any direct, indirect, incidental, or consequential damages arising from your use of this tool. You use it at your own risk.

5. Acceptable Use

You agree not to use this tool for any unlawful purpose, to make financial or military decisions, or to represent AI-generated assessments as authoritative intelligence.

6. Changes

We reserve the right to modify these terms at any time. Continued use constitutes acceptance.

◆ Privacy Policy

Last updated: April 2026

Infrastructure & Cloudflare

This site is hosted on Cloudflare Pages. As your CDN and infrastructure provider, Cloudflare processes every request to this site at the network edge. This means Cloudflare handles standard server-side data including IP addresses, user agent strings, request URLs, and timestamps. This is inherent to how web hosting works and is not unique to Cloudflare. Cloudflare may also provide aggregate, privacy-first Web Analytics (page views, countries, device types, referrers) without cookies or individual user tracking. Cloudflare does not sell personal data. See Cloudflare’s Privacy Policy for full details.

What We Collect

Beyond what Cloudflare processes as infrastructure provider, we collect nothing. The Monitor itself sets no cookies, uses no additional analytics, and does not track, store, or transmit any personal information about visitors.

Local Storage

Assessment history and display state may be cached in your browser’s localStorage to improve load times. This data never leaves your device and contains no personal information.

External Requests

When the assessment refreshes server-side, the backend connects to: Anthropic API (AI analysis), GitHub API, HuggingFace API, Wikimedia API, and jsDelivr CDN (map data). Your browser directly requests map data from jsDelivr on load. Each service has its own privacy policy.

Cookies

The Monitor does not set any cookies. Cloudflare may set functional security cookies (e.g. __cf_bm) as part of its infrastructure services. These are essential for security and do not track you for advertising purposes.