The web’s foundation is shifting. For years, publishers relied on Google’s blue links to bring readers to their sites. But with AI-driven search tools now summarizing information directly in results, that relationship is breaking down.

Some publishers say search referrals have dropped more than 30% since Google’s AI Overviews launched. That change is forcing media companies to decide whether to let AI systems use their content, block them entirely, or negotiate licensing deals.

Recent comments from Reddit’s CEO Steve Huffman and updates from Gannett, parent company of USA Today, offer a clear snapshot of how differently companies are responding to this shift.

Reddit says chatbots aren’t a major traffic source yet

Reddit CEO Steve Huffman told analysts during the company’s Q3 2025 earnings call that AI chatbots are not a significant driver of site traffic. According to Huffman, most users still find Reddit through Google Search or direct visits.

That insight challenges the idea that chatbots are already replacing traditional search. Still, Reddit is preparing for that future. The company has licensed its data to OpenAI for model training and is renegotiating its deal with Google for fairer terms. It’s also merging its core search with Reddit Answers as it plans to become a go-to search engine. 

It’s also developing tools for publishers such as AI-powered link tracking, RSS automation, and recommendation systems to make external media more visible within Reddit’s ecosystem.

These steps show how Reddit is trying to adapt before AI search becomes dominant. The company’s stance shows it wants to benefit from search while also seeking fair terms for how its data is used by AI systems.

Gannett takes a firmer stand against AI crawlers

Gannett, the parent company of USA Today, is taking a more defensive approach. In a LinkedIn post highlighted by Glen Allsopp, Gannett had blocked about 75 million AI bots, including those from OpenAI. The company says it wants Google to display its content as traditional links rather than as part of AI-generated answers.

This position is similar to that of the News/Media Alliance. In May, the group accused Google of “theft” for using AI to repurpose articles without compensation. The publishing group claims that Google’s AI Mode “takes content by force and uses it with no return.”

Gannett is also experimenting. The company is testing DeeperDive, an AI-powered Q&A tool built with Taboola, which lets readers explore summarized stories directly on its own sites. It has also entered partnerships with Perplexity’s Publisher Program and Microsoft’s upcoming Publisher Content Marketplace, both of which offer licensing and revenue-sharing models.

Google and Apple are shaping how AI search reaches users

The dispute goes beyond publishers. Search engines, browser makers, and AI platforms are reshaping the web. According to Bloomberg, Google won’t let publishers fully opt out of having their content used in AI search. Internal documents show that Google discussed giving publishers control over how content is used but dropped the idea, saying AI search is becoming a monetization channel.

Apple is also exploring whether to include AI-first search alternatives like Perplexity in future versions of Safari. That could give iPhone users the option to use Perplexity’s conversational search rather than Google’s default engine.

Infrastructure companies are joining the debate. Cloudflare, which supports a large share of global web traffic, has started blocking AI crawlers by default for new users. It’s also testing a “Pay Per Crawl” feature that would let publishers charge AI systems to access their content.

Rolling Stone’s parent company has sued Google, accusing the search giant of using its material to train AI models without permission. These actions reflect a growing push for compensation and transparency in how AI firms use publisher data.

The new value chain of attention

AI search is quietly redrawing the web’s value chain. The old model of creating content, getting indexed, and earning clicks no longer holds when AI systems can answer questions without sending users back to source sites. 

This shift could alter how audiences discover and engage with content. Publishers that protect their data, negotiate fair licensing, and develop internal AI tools may retain visibility in this new ecosystem. Those that don’t risk becoming invisible to the very audiences they once reached through search.

The takeaway is that the open web is no longer guaranteed to deliver traffic. AI search is becoming the new gatekeeper, and publishers are deciding how to adapt.

‍