Intelligence Disclaimer

Every Praevum assessment is an analytic estimate produced by AI models. It reflects a probabilistic reading of available signal at a point in time — not verified fact, and not a guarantee of any outcome. Confidence levels are calibrated judgments, not statistical certainties.

Assessments are grounded in real, open-source references gathered per request (keyless news and reference material). Cited references are labeled inline (e.g. [S1]) and listed with each response. Each source is graded for reliability — high, medium, or unverified — to help you weigh the evidence. Reliability grades are heuristic and do not certify accuracy.

The system is instructed never to fabricate a source. Where a judgment is the model’s own synthesis rather than directly supported by a cited reference, it is marked as an analytic judgment. Even so, models can err — always verify consequential claims against primary sources.

Praevum outputs must not be used as the sole basis for operational, safety, legal, financial, medical, or life-critical decisions. They are intended to augment the work of qualified analysts and decision-makers, who remain responsible for verification and final judgment.

Situations evolve rapidly. Open-source signal is cached briefly for performance and may lag real-world events. Treat every assessment as a snapshot and re-run analyses as conditions change.

Assessments are provided “as is,” without warranty of accuracy, completeness, or fitness for a particular purpose. Praevum is not liable for decisions made in reliance on its outputs.