Transparency before certainty

How Fhemni reaches an answer.

Fhemni uses AI to analyze public videos and official programmes. Sources, timestamps and assumptions remain visible because AI can miss, misread or overstate evidence.

01

Facts are checked as facts

Claims about existing or past events are compared with sources and labelled supported, contradicted, needing context or unverifiable. A statement in a video is never proof by itself.

02

Promises are forecasts, not facts

A future electoral promise is assessed only for feasibility during the 2026–2031 term. The verdict is not a claim that the promise is true or false and never guarantees what will happen.

03

Chat keeps provenance visible

Video-only answers use the analyzed video context and timestamps. Source-check mode separates what was said from what external evidence supports. An answer is never a new statement by the speaker.

04

No voting recommendation

Fhemni does not represent a party or candidate and does not tell anyone how to vote. It publishes the original material, evidence and uncertainty so readers can inspect the reasoning themselves.

AI-generated analysis can be wrong.

Review the original video or programme, timestamps, sources, data date, model or method information, and revision where available. Publication does not remove uncertainty.

Report an error