Millions of AI-generated pages from Facebook are now ranking in Google Search, turning a quiet corner of the social network into one of the largest machine-written libraries on the open web. The pages live under Facebook’s Answers directory, where each one offers an AI-written summary of a topic built from public Facebook posts. Roughly six million of them are indexed by Google, and they are showing up for millions of queries across regular results, AI Overviews, and AI Mode.
What Facebook Built
Facebook Answers works like a reference library assembled by machines. Each page takes a single topic, writes a summary about it, and leans on posts and pages from Facebook users as its source material. The directory sits at facebook.com/fb-answers, and it has been growing since the summer, when Meta began filling it at industrial scale.
The scale is what makes the story hard to ignore. About six million of these pages now sit in Google’s index, and they rank for millions of search queries. That puts one company’s AI output on a footing few publishers could ever reach, and it happened out of view, with no announcement from Meta about the directory or its purpose.
How the Directory Was Found
The directory came to light through the work of SEO specialist Saul Westbrook, who noticed in June that Meta had quietly indexed millions of these AI-generated pages. Writing about the find this week, he traced the directory’s growth from its first population in July to the more than six million URLs now visible in Google’s index.
Westbrook also tracked the directory’s organic traffic month by month since June, and the trend points upward. The pages are not just indexed; they are being clicked. Search Engine Roundtable covered the discovery on Thursday, and the story spread quickly among search professionals who track how Google treats machine-written content at scale.
Why SEOs Call It Scaled Content Abuse
Google’s own spam policies define scaled content abuse as producing large volumes of pages to manipulate search rankings, regardless of whether the content is written by people or machines. SEO consultant Glenn Gabe, who amplified Westbrook’s findings, said the Facebook directory fits that definition exactly, calling it another example of massive scaled content abuse.
The visibility goes beyond the ten blue links. Gabe noted that the pages are showing up in AI Overviews and AI Mode as well as in Gemini, which means the same AI-written summaries are being recycled as answers inside Google’s own AI products. That is the part of the story that has drawn the sharpest reaction from the search community.
Original Analysis: What This Changes
The uncomfortable question is whether the rules apply equally. Small publishers have been hit hard for publishing AI content at scale, and entire sites have vanished from Google’s index for doing far less. If a directory of that size keeps ranking, it will look like enforcement depends on who you are rather than what you did.
There is a circularity problem that Google has not addressed. AI Overviews and AI Mode are supposed to synthesise the best of the web, but here they surface summaries that were themselves generated by AI from social posts. Each layer of automation makes it harder for a searcher to tell where the information came from, or whether anyone checked it.
For Meta, the directory shows its AI ambitions now extend past chatbots and feeds. The company has woven AI through its products all year, and Answers turns its archive of posts into searchable reference material. Whether that was the plan or an experiment, the result is a new kind of publisher: one that never sleeps and never hires a writer.
There are also reports that Instagram, which Meta owns, may be doing something similar. If that is true, the footprint of AI-generated social content in search could be far larger than the six million pages already counted, and the true scale of the experiment is still unknown.
Barry Schwartz of Search Engine Roundtable has suggested the directory may not stay indexed for long now that it is public. The coming weeks will show whether Google treats it as spam or as a legitimate source, and publishers will be watching the index count closely.
Why It Matters
This is a test of Google’s spam enforcement at a scale no independent publisher could match. If six million AI-generated pages from a tech giant keep ranking, publishers penalised for far smaller operations will reasonably ask what the policy actually forbids.
If Google acts and the directory disappears, it will show that even the largest companies face the same rules. Either outcome shapes how much machine-written content the web’s biggest search engine is willing to carry, and every publisher watching this case will read the result as a signal.
Background
Google has spent the past two years tightening its spam policies against AI-generated content at scale, while its own AI Overviews and AI Mode have pushed more searchers toward summarised answers. Meta, meanwhile, has made AI central to its products, from chatbots to content generation. The Facebook Answers directory sits exactly where those two strategies collide.
Source: Search Engine Roundtable report
0 Comments