A self-contained instrument for two pre-registered behavioral studies on AI-disclosure labels. Every action is logged server-side (write-ahead). The same pages run as the live, data-collecting instrument and as a browser-only preview.
Does an “Created Using Generative AI” label lower readers' endorsement of identical human-written fiction? Archive-style reading page → kudos intent (primary) → behavioral click-through → effort/authenticity mechanism.
Open Study 3 →
Do stigma / benefit regimes change whether writers truthfully label their own AI use? Gallery (regime carrier) → writing with a logged AI assistant → regime revealed → label choice. Truthful disclosure given logged use is the primary outcome.
Open Study 4 →
Researchers: live monitoring dashboard — live cell counts, the Study-4 adaptive AI-user counter, the intake kill switch, and raw-data export (requires the admin key).