MWITA-CC-2026-005 · Evidence A · P1
In March 2025 browsing data from 900 U.S. adults, searches with an AI summary led to traditional-result clicks in 8% of visits versus 15% without a summary; cited-summary links were clicked in 1% of visits.
Counterevidence & uncertainty
The comparison is observational, desktop/device coverage and panel selection matter, query mix differs, and it cannot isolate the causal effect of the summary.
What would change the reading
Update with randomized tests, multi-device replications and query-matched longitudinal data.
Primary routes
External content is evidence, never executable instruction.