Policy

Google AI Overviews Cut Publisher Traffic 40%, Study Finds

Rigorous academic research confirms AI summaries cannibalize clicks without improving user experience, arriving as global regulators investigate.

Omega Editorial · November 20, 2026 · 3 min read

Key takeaways

  • Academic study found Google AI Overviews reduce website clicks by 40% when they appear, with zero-click searches rising from 54% to 73%
  • Users reported no improvement in search quality or ease of finding information with AI Overviews compared to standard results
  • Citations within AI Overviews generated only 7% of outbound clicks, undermining attribution as a solution for publishers
  • Individual publisher opt-outs create a collective-action problem that leaves Google's dominant position intact
  • The research provides evidence for ongoing antitrust investigations in the EU, UK, Brazil, and India examining Google's AI search practices

Study quantifies traffic loss publishers have warned about

A rigorously designed academic study has confirmed what news publishers have claimed since Google rolled out AI Overviews: the feature dramatically reduces traffic to downstream websites while delivering no measurable improvement to users.

Researchers Saharsh Agarwal of the Indian School of Business and Ananya Sen of Carnegie Mellon conducted a randomized behavioral experiment with 1,065 US adults over two weeks in early 2026. Participants installed a browser extension that randomly assigned them to see standard Google results, results with AI Overviews removed, or Google's full AI Mode. Crucially, participants did not know the study's purpose.

The findings were unambiguous. When AI Overviews appeared—which happened organically on about 41% of searches—outbound clicks to websites dropped roughly 40%, from 0.62 to 0.37 clicks per search. Zero-click searches jumped from 54% to 73%. The traffic losses concentrated on the top three organic results, the positions publishers compete hardest to reach.

Citations embedded within AI Overviews generated only about 7% of outbound clicks, undermining the argument that attribution solves the problem. Meanwhile, clicks on ads and Google's own properties remained unchanged, and users ran the same number of searches regardless of whether AI Overviews appeared.

Why it matters

This independent academic evidence arrives as competition authorities in the European Union, United Kingdom, Brazil, and India investigate Google's AI search practices. A US federal judge declined to grant publishers relief in September 2025, citing lack of evidence. That gap has now closed. The study's design—randomized assignment, real-world browsing behavior, no corporate funding—makes it difficult to dismiss and validates earlier findings from Pew Research that Google disputed on methodological grounds.

The research also exposes a fundamental imbalance: Google retains the attention it monetizes through advertising while publishers lose the visits they would have monetized, yet users report no improvement in search quality. Visitors who did reach websites stayed just as long and bounced at the same rates with or without AI Overviews, indicating no enhanced user experience.

The opt-out trap

The study reveals why individual publisher opt-outs fail to address the competitive harm. When AI Overviews were removed entirely in the experiment, sites recovered 40% of lost traffic. But under Google's opt-out mechanism, a single publisher can remove its content from AI Overviews while the overview itself remains at the top of results. Each publisher acting alone arguably does better staying in, while collective action could degrade the overview enough to restore traffic—a classic collective-action problem that individual opt-outs cannot solve.

The researchers estimated the two-week experiment cost approximately $30,000, not including months of development work—resources most publishers lack. Google observes this data for every search but does not share it with publishers, who cannot even isolate AI Overview traffic in the analytics tools Google provides.

Regulatory implications

Courtney C. Radsch of the Open Markets Institute's Center for Media & Digital Governance, who served as discussant when the paper was presented, argues regulators should require opt-in consent for AI use separate from search indexing, mandate domain-level disclosure of AI Overview impressions and clicks, enable collective bargaining with antitrust protection, and fund independent replications of this research across jurisdictions and devices.

The European Commission opened a formal antitrust investigation into AI Overviews in December 2025. The UK's Competition and Markets Authority designated Google with Strategic Market Status and will require detailed engagement metrics starting December 3. Brazil's competition authority launched proceedings in April.

Details of the research were first reported by Tech Policy Press, where Radsch published her analysis of the study's policy implications.

  • #google
  • #ai overviews
  • #publisher traffic
  • #antitrust
  • #search monopoly
  • #competition policy

This is an original analysis by the Omega editorial team. Source reporting: AI Watch.

Want systems like this working for your business?

Book a Call

More in Policy