AI Disclosure

Deterministic systems

Forecast Alpha's probability engine, calibration, risk controls, and execution gates are versioned calculations. Generative AI does not authorize orders or silently promote a model.

Generative systems

AI may classify source relevance and direction, normalize themes and entities, or draft research explanations. The service records the provider, model, prompt version, citations, confidence, and input digest for governed agent proposals.

Data and limits

Public source excerpts and the minimum required research context may be sent to configured AI providers. AI can hallucinate, omit evidence, misread a resolution question, or reflect source bias. Review citations and point-in-time timestamps before relying on output.

Human authority

Agent output is append-only proposed analysis. It cannot alter raw evidence, promote a champion model, or enable capital authority.