The Confidence score reflects how strongly AI models support their dimension ratings with specific, verifiable reasoning. Low confidence often correlates with vague answers, hedging, or admitted lack of knowledge.
Confidence is computed by the AIRepScore platform from model outputs — it is not a subjective user rating. Use it to decide whether a low score reflects genuine weak perception or simply sparse AI knowledge.
Always interpret dimension scores together with Confidence and Consistency.