Data & AI transparency

Why this page exists

A financial product that uses models has to say where. Without that you cannot tell whether a figure on the screen was read from a source or produced by a model — and that difference is the whole difference.

The rule

A model does not produce a financial figure. No value shown as data — yours, official or comparative — is the output of a language model. Data comes from what the user supplied or from what was read from a publication, and passes through deterministic calculation.

Where models are used

• In development — writing code, tests and documentation, under human review.
• In drafting explanations — text that says what a figure means. Such text is labelled BeeWiser editorial and is not an independent source of fact.
• In language processing — organising and cataloguing existing material.

Where they are not

• In setting the value of any financial figure.
• In generating a peer group or a reference value.
• In deciding whether a comparison is valid — that is a rule, not a judgement call.
• In a personal recommendation presented as checked when it was not.

Text we wrote ourselves

An explanation written by us carries its own label. It rests on other sources and does not stand on its own as fact.

Automated decisions

We do not make automated decisions producing legal or similarly significant effects on a person. We do not rate people, do not score them, and do not pass such an assessment to anyone.