When the Veracity Engine detects distorted facts in an AI response, it records each specific misrepresentation in a structured distortions array and penalizes the URL's overall fidelity score.
flowchart LR
A[AI Response Segments] --> C[Veracity Evaluator]
B[Source Page via url_context] --> C
C --> D[Survived Facts]
C --> E[Lost Facts]
C --> F[Distortions Array]
D --> G[Fidelity Score Calculation]
E --> G
F --> G