Context
A preregistered field experiment with 1,100 participants found that removing AI Overviews and AI Mode increased click-through rates to publishers, while an AI Mode-only experience reduced publisher click-through and weakened reported user experience and trust.
What changed
The study provides causal experimental evidence rather than observational traffic correlations, comparing Google Search experiences with and without generative AI features.
Why it matters
For GEO and AI visibility measurement, appearing in an AI answer cannot be assumed to preserve publisher referral traffic. Visibility, citation, referral, and user trust should be measured separately.
Visibility OS interpretation
Visibility OS treats this as strong evidence about downstream traffic effects, not as evidence about how Google selects or ranks sources inside AI answers.
What remains uncertain
How do these effects vary by query type, publisher category, citation prominence, and repeated exposure over longer periods?