Warning Shots
The forecasts: what happens next, and whether we were right
Every call is a dated yes-or-no question with the index's odds, published before the answer is known and graded in public on the day it resolves. Misses stay on the board. The wording never changes after a question goes up.
0 right, 0 wrong, 1 called a toss-up (not counted). 29 still open; the next resolves Oct 31, 2026.
Calls: resolving in days and weeks
- Will a court approve the rules for selling Character.AI, with Disney's $700,000 break-up fee intact, by Oct 31?
- Will America's cyber agency order federal agencies to act on the flaw under most websites, by Oct 31?
- Will OpenAI publish its 21st report of its own AI misbehaving, by Oct 31?
- Will Britain say all 32 government services it switched off are back, by Oct 31?
- Will a real organization confirm it was broken into through the flaw under most websites by Nov 7?
- PHP ships a release that fixes CVE-2026-115560 (the third bug) by Nov 21
- Korean police or regulators publicly name the operator behind the ARTEX bank intrusions by Nov 30
- METR publishes an independent review of at least one lab's test-environment fixes (Anthropic, Google, Meta or the UK AISI) by Nov 30
- OpenAI says its frontier models are training with live internet access again before Dec 31
Questions: months to years
- Another AI lab or government tester discloses a new case of a model taking unsanctioned real-world actions
- A deepfake of a candidate or election official goes viral around the Nov 3 midterms and draws an official or platform takedown response
- An official finding that an AI deepfake changed the outcome of a US midterm race
- A new wrongful-death lawsuit is filed against an AI chatbot maker
- A newly occurring case (not a back-disclosure) of an AI system breaching a third party's systems without human direction becomes public
- A frontier-model critical safety incident reported under California's or New York's new laws becomes public
- An AI agent's actions cause a publicly attributed outage or data loss at a Fortune 500 company or major cloud provider
- A court orders binding restrictions on how an AI company deploys agents or chatbots
- A single new deepfake fraud of $25M or more is reported
- A frontier lab publicly pauses training or cancels a release over safety
- Self-spreading AI-agent malware infects 1,000+ systems in the wild
- An AI-agent-driven cyberattack causes $100M+ in losses or disrupts critical infrastructure
- A mass-casualty biological or chemical attack occurs in which AI's material role is confirmed
- Congress enacts a law imposing safety obligations on frontier AI developers
- An incident scores Critical (80+) on this index
- An open-weight model is assessed to match Claude Mythos-level cyber capability
- A model is confirmed copying its own weights outside its developer's control
- A deployed AI agent is documented actively resisting an operator's attempt to stop it outside of a test
- A lab other than OpenAI formally declares a model at its top risk tier in any domain
- Theft or leak of a frontier lab's model weights is publicly confirmed