Facebook has been quietly publishing AI-generated pages under facebook.com/fb-answers/. Each one is a summary that explains a topic, built from posts by Facebook users and pages. Barry Schwartz reported on 5 November 2026 that Google has indexed more than six million of them, and that they are ranking in Google Search, AI Overviews, AI Mode and Gemini. The pages were first indexed in June 2026 and started filling up in July. Saul Westbrook of Peripheral Creative and Glenn Gabe flagged it first. Neither Google nor Facebook has commented.
Whether Google acts is a side question
The headline asks whether this is scaled content abuse, which is Google’s name for its rule against pages produced in bulk to game search. The pages are ranking while the debate runs.
Two reactions follow the story. The first says Google has to act, because six million auto-generated pages is what the rule was written for. The second says this proves AI content is slop, and that anyone publishing with AI is next in line.
The first is Google’s problem to settle. The second is wrong, and it is the one that will cost operators money if they believe it.
Look at what the pages actually are. Each one summarises what other people posted on Facebook. No author holds a view on the topic. No page makes a claim that a competing page could not make. The pages exist because Facebook has a lot of posts and a model that can summarise them, and that is the whole reason. The word in Google’s rule is scaled, and what got scaled here is a summary with nobody behind it.
Slop comes from how AI is deployed
Slop is what you get when you expect AI to think on its own and mass-produce content with no direction and no opinion. The output comes out generic. Search engines and AI systems are getting better at recognising generic, and they show it less.
I run content at volume myself. This site is one example. Claude does the drafting, the counting and the checking. I make the decisions, and every post carries a position I hold and a frame I chose before the draft existed. That context is something only I can supply, so the output cannot be generic even when there is a lot of it. The tool is the same one anybody can use. AI makes volume free, and volume was never the goal. The frame is the goal, and volume is what you get to do once you have one.
What this looks like for a real business
Take a dental clinic with a small marketing team. Someone on the team has seen that AI can write a page in a minute and wants 400 pages, one for every treatment in every nearby neighbourhood. The spreadsheet is ready.
The question I would ask is the one I ask of any new tool. Is there a need? What does page 212 say that page 211 does not? If the honest answer is that the clinic’s name and a suburb got swapped in, those 400 pages are Facebook Answers at a smaller scale. A summary with nobody behind it, times 400.
Now the same clinic, the same tool, a different approach. The dentist writes down what the clinic actually does differently: which treatments it completes in one visit, what a first consultation costs, which cases it refers out. Each page then answers one specific question a patient asks, with one of those facts in it. The pages can still be produced at volume. Engines quote specific facts and drop adjectives, so the specific fact is what gets the clinic named in the answer.
We do this at AeroChat. It publishes many pages, each one about a single integration with a specific piece of software, each one catching one long-tail question. That is volume. Every page states something only we can state about our own product.
What to do
- Before producing pages at volume, write the frame first: what you are, who you serve, what is measurably different about you. If you cannot write it in one page, the AI cannot write it in 400.
- Give every page one specific claim or number a reader could check. If you cannot find one for a page, do not publish that page.
- Judge the output by results. Check whether the engines name you, and date every check, because AI answers change week to week.
- Ignore the takes that say AI content is finished. The Facebook story is about pages with nobody behind them, and yours do not have to be.