Autonomy Risk Countdown

This project is a public monitor for autonomous systems risk. It estimates when current trends in AI capability, military integration, human oversight, governance, and real-world incidents could converge into a high-risk autonomy threshold where meaningful human control starts to break down.

What the project tracks:
- Frontier AI capability and multi-agent autonomy
- Military and institutional integration of autonomous systems
- Human override capacity and governance enforcement gaps
- Current incidents, source-backed evaluations, and estimated risk horizons

What visitors can see:
- A live countdown to the model's current estimated risk horizon
- A recent incident feed with source links
- A history of past evaluations and how the estimate has shifted over time
- Underlying model notes, indicators, and projection details

How the site is updated:
- The main model is periodically reevaluated using grounded web research
- Incident updates are refreshed separately and added to a running public timeline with source links

Main site:
https://ctrl-c.nl/autonomy-risk/

HTML summary:
https://ctrl-c.nl/autonomy-risk/summary.html