Google’s search business is no longer simply sending readers to websites. With AI Overviews and AI Mode answering more questions directly on the results page, the company has begun reshaping the bargain that powered online publishing for two decades—and publishers are feeling the cost in lost referrals.
For years, websites invested heavily in search-engine optimization because appearing near the top of Google could introduce millions of readers to a publication. That system was imperfect, but it gave publishers a reason to create searchable, useful pages and gave Google a constantly refreshed index of the web. The company’s AI-driven search changes are now upending that long-standing relationship by placing synthesized answers between users and the sources that supplied the information.
Google introduced AI Overviews in 2024, then expanded the idea with AI Mode in May 2025. AI Mode replaces the familiar results experience with a chatbot-like interface capable of producing longer responses assembled from multiple web pages. Google still displays links and citations, but a user who receives a satisfactory summary may have little reason to open them—especially for basic questions, shopping research, travel planning, or quick explanations.
The impact is visible in publisher analytics, although the size of the decline varies by outlet and search topic. Similarweb data cited in industry coverage showed organic search’s share of traffic to The New York Times falling from 44 percent three years earlier to 36.5 percent in April 2025. Other major publishers, including CNN, Business Insider, and HuffPost, also recorded significant year-over-year declines. The shift is especially painful for news organizations, whose advertising income depends on converting visits into page views, subscriptions, and repeat readers.
The numbers are not entirely uniform. Research from Digiday found that AI Overviews appeared for a growing share of queries, while early traffic studies produced mixed results across individual publishers. Other analyses found that the top-ranking result received substantially fewer clicks when an AI summary appeared. That uncertainty makes planning harder: publishers cannot easily tell whether a traffic loss comes from Google’s interface, changing audience habits, social platforms, or the broader migration toward standalone chatbots.
Google’s advice to site owners is largely familiar: maintain original, useful content, ensure pages are crawlable, and monitor performance in Search Console. Its documentation says visits from AI Overviews and AI Mode are counted within the broader web-search reporting rather than separated into a fully transparent traffic stream. That may help publishers measure the overall result, but it does not answer the central business question—whether being cited in an AI answer produces enough visitors to replace the clicks that traditional search once delivered. Google’s own guidance confirms that AI-feature appearances are folded into standard Search Console reporting.
For readers, the change promises faster answers and less searching through pages designed primarily to rank. For publishers, it creates a dangerous possibility: supplying the reporting, reviews, tutorials, and expertise that train or feed an answer engine without receiving a sustainable audience in return. The next phase of online publishing will likely involve a mix of subscriptions, direct-reader relationships, licensing negotiations, and attempts to make original work valuable enough that an AI summary cannot replace the source.
