Google once served as the indispensable gateway for online publishers. Traffic flowed freely from its search results. Revenue followed. But that relationship has frayed. Now some of the web’s biggest names, including Reddit, are weighing whether to block the tech giant from tapping their content for artificial intelligence tools.
The shift marks a stark reversal. For years publishers optimized for Google visibility. They accepted the trade-off of free content in exchange for clicks. Yet AI-powered summaries change the equation. They deliver answers directly on Google’s page. Users stay put. Referrals plummet.
Reddit finds itself at the center of this tension. The company signed a deal in 2024 worth about $60 million a year. It let Google train AI models on Reddit’s vast archive of user discussions. Renewal talks are underway. But executives question the value. Google’s AI features often answer queries without sending visitors back to the forum, according to people familiar with the discussions reported by The Wall Street Journal.
So Reddit has explored shutting off access. The move could limit how Google’s models draw on real-time human conversations. Those threads once powered better search results. Now they risk fueling a competitor that keeps users on Google.
Other outlets express similar doubts. USA Today, Politico, Reuters, The Economist and People Inc. all evaluate their ties to the search leader. Traffic data tells a grim story. USA Today’s organic Google referrals fell nearly 50 percent over the 12 months ending in June 2026. Politico dropped 23 percent. Business Insider lost more than 85 percent. The figures come from Semrush and were highlighted in reporting by the New York Post.
USA Today chief executive Mike Reed didn’t mince words. “It’s time to take a stand and say enough is enough,” he told the New York Post. The company already sued Google in 2023 over claims of monopolizing digital advertising. Now it considers blocking AI training access too.
Politico eyes restrictions on its non-paywalled articles. Reuters weighs the trade-off between search visibility and AI data use. The Economist and others follow suit. Their frustration boils down to one point. AI overviews summarize content so thoroughly that clicks evaporate. A Pew Research Center study found users clicked traditional links in 15 percent of searches without AI summaries. That rate fell to 8 percent when overviews appeared. Inside the summaries themselves, clicks registered at just 1 percent.
Google pushes back. The company says it still sends billions of clicks to websites each week. It offers technical controls like the Google-Extended bot token for opting out of AI training. Publishers can also use noindex or nosnippet tags. And in Britain, regulators ordered Google to let sites block AI use without harming their search rankings. The U.K. Competition and Markets Authority ruling could influence global practices.
Yet many publishers remain unconvinced. A class-action lawsuit filed by a group of major publishers and author Scott Turow accuses Google of using millions of copyrighted books without permission to train its Gemini models. The case adds legal pressure atop the commercial strain.
Reddit’s stock reflected the uncertainty. Shares fell as much as 9 percent after the reports surfaced, according to real-time trading data noted across outlets including the gHacks Tech News. The reaction showed how intertwined the companies’ fortunes remain. Reddit still relies on Google for a sizable chunk of its traffic. Cutting off the crawler could hurt both sides.
And here’s the paradox. The very content that made Reddit valuable for AI training came from millions of users drawn in part by Google search. Block the data. Lose the revenue. Keep feeding it. Watch traffic erode anyway. Negotiators face a difficult balancing act as the current agreement nears its end.
TechRepublic captured the broader stakes in its coverage. “The question is no longer whether publishers should appear in AI-powered search,” the publication wrote. “It’s whether the traditional bargain — content in exchange for traffic — still works when AI increasingly keeps users on Google’s own platform.” The July 23, 2026, article drew from multiple reports and noted Google’s claim that its AI features drive additional engagement.
But data from recent months suggests otherwise for many sites. Cloudflare’s chief executive has warned about the long-term effects of reduced bot traffic and referral declines. Publishers experiment with paywalls, subscriptions and alternative discovery channels. Some limit indexing for AI while preserving traditional search. The tactics vary. The underlying anxiety does not.
Google expanded its AI overviews aggressively in recent months. The feature now appears in far more queries. It pulls from licensed data, public web content and partnerships. Reddit’s real-time conversations provided a rich training source. That human element helped models sound natural. Yet when those models answer questions outright, the original forum threads lose their audience.
So talks continue. Google reportedly pitches deeper AI integration in exchange for continued access. Some publishers test partial blocks. Others demand higher payments. The $60 million annual figure that seemed generous before AI summaries proliferated now looks insufficient to offset lost ad revenue and direct traffic.
Industry watchers point to a fundamental power shift. Search engines no longer act solely as neutral directories. They synthesize information. They compete with the sources they once promoted. This evolution forces hard choices on companies built around web traffic.
Reddit’s situation stands out. Its user-generated content proved especially useful for training. Conversations, opinions, troubleshooting threads — all offered nuance that static articles often lack. The 2024 deal helped fund operations ahead of its public listing. But post-IPO priorities have changed. Growth metrics now include user retention and direct visits. AI summaries threaten both.
Similar calculations play out at legacy newsrooms. Politico built its brand on fast, inside-the-Beltway reporting. When Google summarizes a scoop in an overview box, the incentive to click through fades. Reuters values its global wire service reputation. Yet AI aggregation risks commoditizing that work. The Economist’s paywalled model offers some protection. Still, its executives monitor referral trends closely.
Google maintains it pays more than 200 publishers for AI-related licenses. The company argues its tools surface content that might otherwise stay buried. Early tests showed some uplift in certain categories. But the aggregate numbers for major publishers tell a different tale. Declines of 20, 50 or 80 percent compound over time. Advertising rates fall with traffic. Newsroom budgets shrink.
Legal and regulatory fronts add complexity. The U.K. order on opt-outs without ranking penalties could set a precedent. European privacy rules already constrain data use. In the United States, ongoing antitrust cases against Google examine its search dominance and ad practices. The publishers’ copyright suit over book training data represents another vector of pressure.
Meanwhile, Reddit users voice opinions on the platform itself. Some fear reduced visibility. Others support protecting community discussions from unchecked AI scraping. The debate spills into threads about AI, search and media. It mirrors the larger industry conversation.
What comes next remains unclear. Renewal terms could include stricter controls on how Google displays Reddit-sourced answers. Higher fees might compensate for traffic loss. Or the partnership could fracture. If Reddit blocks AI access, Google might rely more on other sources. Models could grow less attuned to current internet slang, niche topics or rapid cultural shifts that Reddit captures first.
Publishers watch each other’s moves. A coordinated approach might strengthen their position. But incentives differ. Some depend more on search than others. News sites with strong brands and loyal audiences may fare better than forums or aggregators. The outcome will shape digital publishing for years.
One thing looks certain. The era of unreserved cooperation with Google search has ended. AI changed the incentives. Publishers respond by asserting control over their data. They question the old bargain. And they search for new paths to reach audiences without surrendering their content’s value.
WebProNews is an iEntry Publication