For two decades, the arrangement between search engines and the people who write for the web was simple. Indie publishers allowed their work to be crawled and indexed. Search engines, in return, sent readers to the source. That arrangement is coming apart, and the people losing the most are the ones with the least leverage: independent publishers, specialist sites and individual writers.
The search industry says it is cleaning up the web, purging the flood of low-quality, machine-generated pages and ‘Ai slop’ that artificial intelligence has made cheap to produce, what some call ai slop. There is truth in that. But it is an incomplete account, and an awkward one, coming from companies that are among the largest investors in and promoters of AI, and whose executives have said openly that the technology will replace much of the work writers do.
Not a Ban, but a Bad Outcome
Some indie publishers have alleged that search engines such as Bing, DuckDuckGo and Google are blocking both original and partially ai-generated content, essentially branding it as ‘Ai slop’, to favor of similar content published by big media. However, at the moment, there is little evidence that Google or its competitors are blocking or even “shadow banning” independent sites through deliberate, manual targeting. What publishers describe is the cumulative effect of ranking updates, spam policies and AI-generated answers placed above the links. For many, the result is indistinguishable from exclusion.
Nate Hake, founder of the travel site Travel Lemming, told the Federal Trade Commission in 2025 that his organic search traffic had fallen 95 percent. Semrush, an SEO ranking service, generated data cited by The Wall Street Journal this summer showed search traffic falling more than 40 percent for some publications over the past year, even as The Guardian and the BBC gained. Penske Media, owner of Rolling Stone, Billboard and The Hollywood Reporter, sued Google in 2025, saying its affiliate revenue had fallen by more than a third from its peak. Google called the claims “meritless” and said last year that total organic clicks had been “relatively stable.” The results suggest that whatever modeling search engines use, it is not good for search in general, whether the publisher is widely-known or not, if the publisher is independent of big media.
A Decade of Filters
The restraints did not arrive all at once. Google’s Panda update targeted “content farms” in 2011, and Penguin went after manipulative links in 2012. The most consequential change for small publishers came in 2022 with the “helpful content” system. Unlike earlier filters that judged individual pages, it classifies whole sites. Google describes it as “entirely automated, using a machine-learning model” that “runs continuously,” and says a flagged site may see the signal applied “over a period of months,” which can weigh on the whole site, though Google says helpful pages on a flagged site can still rank. In March 2024, Google folded the system into its core ranking algorithm. How does a legitimate site differentiate itself from ‘Ai slop’ “content farms” when they’re using ai to aide in content creation? For an independent site, a single verdict, delivered without notice or explanation, can depress an entire domain.
Google has long said it does not use a general “domain authority” score. But internal documentation that leaked in May 2024, which Google confirmed was authentic, referred to an attribute called “siteAuthority.” Google warned against conclusions drawn from “out-of-context, outdated, or incomplete information,” and the documents do not show how signals are weighted. Still, for publishers routinely outranked by established brands on subjects they knew better, the leak suggested that a site’s size and reputation may count for more than Google had acknowledged.
Nor is there much of an exit. Google handles nearly 90 percent of U.S. searches, according to StatCounter, and many of its apparent rivals are not independent in any meaningful sense. Startpage serves results licensed from Google. Yahoo and DuckDuckGo rely mainly on Microsoft’s Bing. Building an index of the web is enormously expensive, so most alternatives rent one from the two companies that have it. For a publisher demoted by Google, the choice of search engines is largely illusory. A federal judge ruled in 2024 that Google is a monopolist in search and in 2025 ordered it to share some index data with qualified competitors. Whether that produces genuine alternatives remains to be seen.
AI Overviews and the Case Against ‘Ai Slop’
Google’s stated rationale is quality. After its March 2024 update, it said users would see 45 percent less “low quality, unoriginal content,” and its spam policies cite using generative AI “to generate many pages without adding value for users.” Meanwhile, AI-generated material now occupies the most valuable position on the results page. AI Overviews and AI Mode summarize answers drawn from publishers’ work, often removing the reason to click; Penske said about 20 percent of Google searches linking to its sites display an AI Overview. Google’s line between AI that manipulates rankings and AI that helps users is defensible in principle. In practice, it makes the company the arbiter of which machine-written text deserves the top of the page, and its own products occupy that space.
Again, it is worth pointing out that alternatives such as DuckDuckGo and Yahoo draw their results largely from Microsoft’s Bing, and Bing, with a small fraction of the market, is the only other index of comparable scale. For independent publishers, that leaves the choice of search engines almost nonexistent.
Where the Money Is Going
The investment behind those products is difficult to overstate. Alphabet reported $51.4 billion in capital spending in the third quarter of 2026, more than double a year earlier, and has guided to $195 billion to $205 billion for the year, driven largely by AI infrastructure. Google also holds a large stake in the AI developer Anthropic, and Microsoft, which runs Bing, owns roughly 27 percent of OpenAI.
Private capital has followed. AI drew more than half of the $512 billion in global venture funding in 2025, according to PitchBook, and an investment partnership that includes BlackRock’s Global Infrastructure Partners agreed last year to buy Aligned Data Centers for about $40 billion. Governments have made AI industrial policy. In January 2025, President Trump announced Stargate at the White House, a venture whose backers pledged up to $500 billion for American AI infrastructure. The European Commission followed with InvestAI, a 200 billion euro initiative.
Writers on Notice
The people directing this capital have been candid about what they expect it to do to work. Mustafa Suleyman, chief executive of Microsoft AI, told the Financial Times in February that most white-collar tasks “will be fully automated by an AI within the next 12 to 18 months.” Dario Amodei, chief executive of Anthropic, warned in 2025 that AI could eliminate half of all entry-level white-collar jobs within five years.
Writers have taken them at their word. In a University of Cambridge study published last November, 51 percent of British novelists surveyed said AI was likely to replace their work entirely, and 39 percent said their income had already fallen. The novelist Tracy Chevalier told the researchers, “I worry that a book industry driven mainly by profit will be tempted to use AI more and more to generate books,” likening it to the way consumers buy “machine-made jumpers rather than the more expensive hand-knitted ones.”
Hollywood fought the question out in public. The Writers Guild of America’s five-month strike in 2023 ended with a contract providing that AI-generated material cannot be considered literary material and that studios cannot require writers to use AI. Adam Conover, who served on the guild’s negotiating committee, said last week that the real danger was never that AI would replace writers, but that the studios “were going to pretend that they had replaced us with AI.”
Newsrooms are already using AI, though most stop short of letting it write the stories. The Associated Press has published automated earnings reports since 2014. The New York Times permits AI for tasks like summaries and headline suggestions but prohibits using it to “draft or significantly revise articles.” Axios has gone further. Under a 2025 partnership with OpenAI, which lets OpenAI train on Axios’s published coverage, it encourages reporters to use ChatGPT. In March, Axios laid off 11 newsroom employees, and its chief executive, Jim VandeHei, told staff that companies “will be structured and staffed differently in the future as tasks and functions are automated,” adding, “Axios is no different.”
A Conflict Worth Naming
None of this amounts to a conspiracy, and it should not be described as one. But it is a conflict of interest, and it deserves to be named plainly. The companies that control how readers find writing are investing hundreds of billions of dollars in systems their own leaders say will do that writing. They are, at once, the gatekeepers of human work online and the financiers of its replacement. When the same firms that warn against machine-made content place machine-made answers above every link, the public is entitled to ask whose interests the ranking serves.
Popdecay is an independent publisher, and that interest should be disclosed. But the consequences extend well beyond any single outlet. When search rewards scale over expertise and summaries over sources, the first voices to fade are those with the smallest legal budgets and narrowest margins: niche critics, regional reporters and writers from communities that larger outlets have historically overlooked.
Want Popdecay without the middleman? Bookmark popdecay.us and come straight to the source.