No. 99
2026-10-26
the AI pulse — brewed, not raw
 AI CRITIQUE
Bir kişinin elleri, parlak turkuaz bir arka plan üzerinde, sayısal verilerle dolu pikselli bir gazete tutmaktadır.

Google's AI search features upend online publishing

Read the article    English (automatic translation)

Summary

Google's AI Overviews (2024) and AI Mode (2025) features have significantly reduced search-driven traffic to online publishers. Some outlets have reported click-through rate drops of up to 89%. This situation is forcing publishers to pivot toward newsletters, social media, and subscription models. Despite this, Google argues that overall click volume has remained 'relatively stable' and that the 'web remains vibrant.'

Why it matters

The article documents a structural shift in how information is consumed. The publisher-Google traffic exchange that has sustained digital journalism for two decades is being replaced by AI summaries that do not direct users to source sites. This threatens the economic sustainability of independent and even major newsrooms. The dismissal of multiple lawsuits indicates that publishers currently lack a legal remedy. The debate extends beyond a single company's product roadmap to concern the general health of the open web.

Highlights

  • DMG Media reported click-through rate drops of up to 89% between June and July 2025. The NYT's search traffic declined from 44% to 35.5% over three years.
  • Nineteen of the 25 member outlets of the DCN (including Bloomberg, AP, Condé Nast, and Vox Media) recorded a 1-25% loss in Google traffic during the May-June 2025 period.
  • The opt-out switch offered by Google for AI Mode/Overviews was widely implemented only after a regulatory order from the UK. Multiple US antitrust and class-action lawsuits against Google were dismissed.

Critical take

The article heavily relies on publisher-side data (Similarweb, DCN, DMG Media). It nearly verbatim repeats Google's counter-claims. It does not question whether 'relatively stable' click volume implies a redistribution of traffic from editorial sites to forums, video, and user-generated content. This nuance suggests the 'disaster' framing may be less universal than it appears. Additionally, the article cuts off mid-sentence. The 'next steps' analysis is lacking, and the actual effectiveness of the opt-out switch is not examined.

Bu kritik, yerel yapay zeka modeli Qwen3.8-27B tarafından yazılmıştır. Haberi ben seçtim; kritik metnini yayımlanmadan önce tek tek denetlemiyorum.This critique was written by the local AI model Qwen3.8-27B. I chose the story; I do not review each critique before it is published.
Back to the issue