ChatGPT Access Cuts Traditional Search by 9%, Shifts Web Traffic

Image credit: Search Engine Journal

U.S. internet users significantly reduced traditional search engine queries following broader access to OpenAI‘s ChatGPT, also altering web referral patterns, according to new research.

Weekly traditional search queries decreased by 9.4 percent with ChatGPT access, a figure that climbed to 17 percent after 20 weeks, raising questions about future content monetization strategies for many websites.

A study by Bocconi University and Comscore found that ChatGPT directs users to external websites in 5.2 percent of sessions, a sharp contrast to Google searches, which generate referrals in 31.1 percent of instances.

The research indicated that ChatGPT referrals primarily target reference, knowledge, tools, SaaS, academic and developer sites.

Ad-supported websites received 27.6 percentage points less traffic from ChatGPT compared to Google, the study reported.

Informational content experienced the most significant decline in traditional searches, with academic research referrals dropping 32.8 percent and general references falling 26.5 percent.

OpenAI gradually expanded access to its generative AI chatbot, initially to paid subscribers in October 2024, followed by free users in December 2024, and finally to anonymous users by February 2025.

The Pew Research Center previously highlighted the increasing reliance on AI tools for information gathering, a trend amplified by the findings.

Companies like Google, YouTube, Reddit and Wikipedia, which traditionally dominate search results and referral traffic, could face substantial shifts in user engagement and advertising revenue, according to analysts.

The study did not specify the long-term financial implications for publishers reliant on ad revenue from search engine referrals.

Source: Search Engine Journal