Picture the scroll. 11:40 p.m., somewhere in Echo Park, a thumb moving on autopilot through Reels. A glossy video slides past: a too-smooth face, a too-perfect latte, a voiceover that sounds stitched together from other voiceovers. For years, the only thing that viewer could do was keep scrolling. Now they can do something much worse for you. They can tell Facebook to stop showing it to them.
Meta launched “Your Algorithm” on Facebook in October, a page in the three-dot menu where people see a summary of what the feed thinks they like and edit it in plain language, for Feed and Reels. On Nov. 10, Meta’s engineering team published how it actually works, and one line should end a lot of content strategies: quality requests, like reducing AI-generated content, “map to dedicated demotion signals rather than topic matching.”
Translate that out of engineer. When someone asks for less AI content, Meta doesn’t just hide a topic. It routes the request into signals built to push content down. And when someone asks to see less of a specific creator, Meta says that creator’s content gets filtered out directly. Not deprioritized. Filtered.
The machinery behind it is not a toy. Meta’s write-up describes a one-sentence “Lately” summary of what you’ve been into, an “Interested In” list prefilled from your behavior, and a “Not Interested In” list fed by every post you’ve hidden or X’d. People can type what they want in their own words. A smaller language model checks each edit, and an LLM reranker does a final pass before anything reaches the screen. This is infrastructure, built to listen to irritation at scale.
Your audience just got a mute button. Your brand is a creator.
For two years the pitch on AI content has been arithmetic. Why make four posts when you can make forty? Why book a photographer when a model hands you a hundred variants by lunch? The cost per asset went to roughly zero, and a lot of brands treated that like a discovery instead of a warning.
Here’s the line item nobody put in the spreadsheet. Every one of those forty posts is now an invitation for your own audience to tell the platform it doesn’t want you. The volume play used to cost you attention. Now it can cost you distribution, and distribution is the whole damn game.
Think about what “less of this creator” means for a brand account. A human creator who annoys you once might get forgiven next week, because there’s a face and a personality to come back to. A brand account that annoys you with its fourteenth synthetic carousel this month has no such credit. It was never a relationship. It was a delivery schedule. And people cancel delivery schedules without a second thought.
Meta isn’t alone, and it isn’t first. Pinterest started labeling AI-generated and AI-modified Pins in April 2025 and promised a “see fewer” option for GenAI Pins, starting in categories like beauty and art. By October 2025, after app-store reviewers complained that AI Pins were “cluttering up Pinterest,” it let people dial generative AI content down across categories. YouTube tightened its Partner Program rules on July 15, 2025, to better identify mass-produced and repetitious content, a change its creator liaison Rene Ritchie called minor. Minor for YouTube. Not minor for anyone whose channel was a content mill.
Notice the pattern. The platforms that spent years rewarding volume are now building tools that let the audience punish it. They aren’t doing this out of taste. They’re doing it because feeds full of slop are bad for time spent, and time spent is the only product they actually sell.
And audiences are primed. Marketing Dive, citing a 2026 Gartner study, reported that half of consumers prefer brands that don’t use generative AI in their advertising. Pinterest itself has claimed that more than half of all online material is now made with generative AI. Put those side by side: supply is exploding while half the demand is looking for the exit.
It isn’t only the audience souring on volume. A Harvard Business Review study by Syracuse professor Adam Peruta found that AI-generated video ads performed worse than human-made ones on both short-term sales potential and long-term brand equity, even when viewers couldn’t tell which was which. So the slop doesn’t just risk a mute. Even the pieces that sneak through are pulling less weight than the work they replaced.
Now the honest counterargument, because it’s a good one. Most people will never open this menu. Defaults rule the internet. The share of Facebook users who will lovingly curate a “Not Interested In” list is probably small, and the feed still responds to what people actually watch. So the risk is overblown, right?
Wrong, and here’s why. The people who bother to use these controls aren’t random. They’re the most engaged, most opinionated, most culturally plugged-in users on the platform, which is to say the exact people a brand has to win before anyone else will care. You don’t lose the passive scroller first. You lose the tastemaker. And when a creator mute filters you out directly, losing that person isn’t a dip in engagement. It’s a door closing, quietly, with no notification.
There’s a design principle buried in Meta’s own write-up that deserves a frame. When someone adds a new interest, Meta’s engineers want it to appear in the feed without being allowed to take the feed over. That’s their rule for the algorithm. It should be yours for the content calendar. Show up with something worth the slot. Don’t flood the zone and pray.
So what does good look like now? Fewer things, made with intent. One real shoot with a real person in a real place, say a chef in Highland Park plating at 4 p.m. before service, captured once and cut five honest ways instead of fifty synthetic ones. Use AI to cut, caption, resize and test. Don’t use it to manufacture a fake life for your brand at industrial scale.
This is where culture does the heavy lifting. Think about the accounts in Los Angeles that people actually defend: the taco truck with one camera-shy owner and a line around the block, the Koreatown bar whose entire Instagram is the same corner booth on different nights, the skate shop in Venice that posts twice a month and still sells out drops. None of those accounts are optimizing output. All of them are building the kind of attachment that makes “see less of this creator” unthinkable.
If you run a brand account, the audit takes an afternoon. Pull the last 90 days of posts. Mark which ones a human actually shot, wrote or performed, and which ones were generated or templated into existence. Then compare saves, shares and real comments, the signals that mean someone wanted more. Our bet is you’ll find a handful of posts carrying the account and a long tail of filler quietly teaching the feed that nobody cares.
We use these tools every day, and we aren’t precious about it. But we’ve stopped judging a month by post count, because the number that matters now is how many people would choose to see more of you if the platform asked them. The platforms are now asking. Some of your followers have already answered.
The slop era ran on one promise: more is free. That promise just got a price tag, and your audience is the one holding the scanner.
