AI Drained the Internet. Now It’s Paying the Price

When OpenAI scraped the internet to train the first version of ChatGPT, the web was an open pantry — years of blogs, books, and Reddit threads laid out for the taking. The company and its peers feasted. But what once looked like a limitless buffet now resembles a barricade. Publishers are suing, content platforms are charging, and the web — that messy, generative chaos of human thought — is closing its doors.

That era has ended.

Publishers are no longer quietly watching their work repurposed by large language models. They’re demanding payment for what was once free. Cloudflare, which manages about a fifth of the internet’s traffic, has become a gatekeeper, blocking AI crawlers by default unless they pay. The high-quality data needed to train the next generation of models — journalism, academic papers, cultural archives — is drying up. In a bitter twist, AI companies may have poisoned their own well.

The numbers tell the story. According to Cloudflare’s analysis, Anthropic’s Claude made nearly 71,000 page requests for every single referral it sent back to publishers. OpenAI’s ratio was 1,600 to one; Perplexity, 200 to one. Google, long accused of siphoning media traffic, looks almost charitable by comparison, with a 9-to-1 ratio. For all the tension between publishers and Google, that relationship — content in exchange for visibility — now feels quaint, even fair.

Since Google introduced its AI Overviews feature in May 2024, the proportion of news searches that result in zero clicks to publisher websites has soared from 56% to nearly 69%. Organic traffic to news sites collapsed from 2.3 billion visits at its peak to under 1.7 billion by May 2025. The shock has gutted an already struggling media industry: layoffs, hiring freezes, and the looming threat of bankruptcy for outlets that once depended on clicks to survive.

It’s a feedback loop with dire consequences. As publishers produce less — or lock their archives behind paywalls — AI companies lose access to the fresh, factual material their systems rely on. That, in turn, means the web’s content becomes increasingly stale and synthetic, dominated by machine-written filler that echoes itself. The internet, once a collective act of human creativity, risks becoming a hall of mirrors.

The Licensing Gold Rush Begins

Lawsuits have already begun to reshape the economics of the web. Anthropic recently agreed to pay at least $1.5 billion to settle a class-action lawsuit from authors — roughly $3,000 per book. Reddit, which struck $203 million worth of licensing deals in early 2024, is negotiating even more lucrative agreements with Google and OpenAI, exploring “dynamic pricing” that would charge AI companies more as the value of its user data rises.

Even platforms built on the ideal of open information are retreating. WalletHub, a financial advice site, recently pulled 40,000 pages of public content behind a login wall. Its CEO, Odysseas Papadimitriou, likened dealing with AI firms to being shaken down. “Either they shut down the road your restaurant is on and no customers can reach you,” he said, “or they keep the road clear but open a restaurant next door to yours and make you serve their customers for free.”

Cloudflare’s move to block AI crawlers by default marks a decisive shift. “The deal that Google made to take content in exchange for sending you traffic just doesn’t make sense anymore,” CEO Matthew Prince wrote. The implication: AI companies have taken far more than they’ve given back.

What Comes Next

In response, the industry is trying to impose order. A consortium of publishers and tech companies — including O’Reilly Media, Reddit, Yahoo, and Medium — has proposed a new protocol called Really Simple Licensing (RSL). The system would require AI crawlers to present valid license tokens before accessing content, much like a digital passport for data.

But even that effort may be too little, too late. As journalist and media critic Pete Pachal notes, there are “myriad” ways for AI firms to bypass blocks — relays, proxy systems, bot disguises — making enforcement feel like an endless “whack-a-mole” game.

Google, meanwhile, finds itself trapped in a contradiction. In a recent court filing, it admitted that “the open web is already in rapid decline,” even as executives publicly insist the web is “thriving.” The confession lays bare what many technologists already suspect: the open web, once defined by reciprocity and discovery, is dying under the weight of generative AI.

The golden age of “free” data is over. The next wave of artificial intelligence won’t just depend on smarter algorithms — it will depend on checkbooks. AI companies will have to start paying for what they once took. The only question now is how much, and whether the humans who built the internet will finally be compensated for feeding the machines that replaced the

Stay ahead in the world of AI, business, and technology by visiting Impact AI News for the latest news and insights that drive global change.

Got a story to share? Pitch it to us at info@impactnews-wire.com and reach the right audience worldwide!