Intelligence

ai

AI in Search Reduces Publisher Referrals Without Improving User Experience: Experimental Evidence

Source
arXiv — Computers and Society
Published
Last verified
20 Aug 2026
Confidence
High
Evidence
Original document retained
Reading time
1 min
Country
International
Relevant to
Research & Evidence, Technology & Data

Executive summary

What happened, and why should leadership care?

Experimental research indicates that the integration of generative AI features, specifically 'AI Overviews' and 'AI Mode,' into a dominant online search platform significantly reduces user click-through rates to original publishers. This reduction occurs without a corresponding improvement in user experience or trust in the information provided. The study suggests that AI-synthesized answers divert traffic from content creators, potentially impacting their visibility and revenue models.

Why this matters

Why is this strategically important?

The findings underscore a potential systemic shift in information consumption and content monetization models. Organisations relying on organic search traffic for visibility, engagement, or revenue generation may face significant headwinds as AI integration in search becomes more prevalent. This necessitates re-evaluation of digital strategy and engagement channels.

Key insights

What should be noted from the evidence?

  • Integrating generative AI into web search changes how users navigate and assess information.
  • A preregistered field experiment on a dominant search platform (N=1,100) measured the causal effects of AI Overviews and AI Mode.
  • Removing AI Overviews and AI Mode increased click-through rates to publishers.
  • An 'AI Mode-only' experience reduced click-through rates and negatively impacted user experience.
  • User trust in information found on the platform was eroded by the AI Mode-only experience.

Evidence and confidence

How far can this assessment be trusted?

High confidence. Named institution, original document retained and analysis corroborated.

Analysis is prepared editorially by Aziz Shuaib Ausi. The original publication remains the authoritative record, and executive judgement remains entirely human.

Source

Where does this originate?

Reported by arXiv — Computers and Society · International. This briefing summarises the publication for executive use; the document itself is not reproduced here.

Read the original publication