Most AI policy debate happens in the abstract: would a binding international treaty help, would a six-month training pause help, would mandatory interpretability help? The AI Consciousness Tracker turns those abstractions into modellable interventions, each with explicit effect estimates, and lets you see what they project onto the Consciousness Index and the eight threat vectors.
The scenario tools are not predictions. They are a structured way to make your assumptions explicit, run them through a consistent causal matrix, and compare outcomes across alternatives.
OPEN THE COUNTERFACTUAL LAB →Pick one of the preset interventions or design your own. The Tracker computes:
The Tracker ships with ten carefully-calibrated preset interventions. Each was researched against published policy analysis and effect estimates were assigned to be realistic rather than aspirational.
If none of the presets match what you want to model, the Custom Intervention Designer lets you define your own: name it, write a description, set per-factor effect sliders (−30 to +30), per-threat effect sliders, and a feasibility score. Custom interventions persist in your browser's localStorage alongside the presets and feed the same computation pipeline.
This matters because preset interventions encode an editorial view of what the relevant interventions are. The custom designer is the escape hatch: if you think the relevant intervention is something we have not anticipated, you can model it and compare.
DESIGN YOUR OWN INTERVENTION →The Scenario Lab is the longer-form sibling of the counterfactual tool. Save full snapshots of your CI factor settings under a name (“baseline”, “optimistic 2030”, “moratorium succeeds”), come back later, and compare them side-by-side. Snapshots persist locally and can be exported as JSON for sharing or version control.
The Time Machine is the historical complement: scrub from 2020-06 through the present using fourteen recorded snapshots of the actual values to see how the Consciousness Index, factor levels, and threat vectors evolved. Press PLAY to scrub continuously and watch the curve accelerate. Useful for seeing how rapidly the present configuration assembled itself.
OPEN THE TIME MACHINE →Effect estimates for preset interventions are editorial judgments, calibrated against published policy literature where possible and labeled as estimates where not. Custom intervention effects are entirely user-defined and carry no calibration; treat their projections as thought experiments, not forecasts. The underlying causal influence matrix between threats is documented in the Tracker's methodology panel.