Across the digital commons, four of the internet's most-used platforms have begun drawing a quiet but consequential line between human expression and machine-generated imitation. Snapchat, YouTube, LinkedIn, and Substack are each, in their own way, choosing to deprioritize wholly synthetic content — what the industry now calls 'AI slop' — not by banning the tools that create it, but by refusing to amplify the flood it produces. The move reflects a deeper anxiety: that when fakeness is rewarded at scale, trust in everything erodes, and the shared sense of reality that makes communication meanin
Major platforms unite against 'AI slop' with content authenticity crackdowns
Platforms that reward fakeness will create a race to the bottom.
Mark
Why now? AI tools have existed for a while. What changed?
Mimi
The scale changed. It became trivial to produce thousands of pieces of content automatically. Platforms started seeing entire channels and accounts that were nothing but machine-generated material, some making serious money. At a certain point, the problem became visible enough that it threatened the platforms' own credibility.
Mark
But these companies built the AI tools in the first place. Aren't they partly responsible?
Mimi
Absolutely. Snapchat, YouTube, and the others all offer their own AI features. They're not saying AI is bad. They're saying that content which is entirely synthetic, with no human judgment or creativity involved, doesn't belong in the feed. There's a difference between using AI to help you write something and using AI to generate thousands of identical posts.
Mark
What about the creators who actually use these tools legitimately?
Mimi
That's why the platforms are being careful about the distinction. LinkedIn removed the automatic "enhance with AI" prompt, but they're not punishing people who use AI to refine their work. YouTube is blocking generic, repetitive, template-based videos from monetization—the stuff that's obviously farmed. If you're a real creator using AI as a tool, you should be fine.
Mark
Is this actually going to work?
Mimi
It's a start. The real test is whether users trust the platforms again. Research shows that the more fake content people see, the less they believe anything. These moves are about rebuilding that trust. But it requires constant vigilance. The moment a platform relaxes, the spam comes back.
Mark
What happens to all the AI slop that's already out there?
Mimi
It doesn't disappear. It just stops getting recommended. It stops making money. Over time, the incentive to create it diminishes. But the internet doesn't forget—that content is still there, just less visible.
O Pulso
- Up to 40% of social media writing may now be AI-generated, and users exposed to synthetic content grow increasingly unable to trust anything they encounter online.
- LinkedIn alone has blocked billions of AI-generated comment attempts in recent months, adding a public reporting button this week as the volume of automated posts reaches hundreds of thousands per day.
- YouTube is cutting off monetization for generic, repetitive, and template-based videos — the exact formats generative AI produces fastest — after discovering entire channels built on synthetic content earning millions in ad revenue.
- Snapchat is removing wholly AI-generated videos from its Spotlight recommendations, while preserving AI-assisted human content, drawing a deliberate distinction between augmentation and replacement.
- Substack's CEO warned that platforms rewarding fakeness will trigger a race to the bottom, launching a tool to help readers identify AI-written text as the platform repositions authenticity as a competitive value.
- None of the four platforms have banned AI outright — the collective bet is that surfacing human-made work, not eliminating machine tools, is what will sustain long-term audience trust.
Across the digital commons, four of the internet's most-used platforms have begun drawing a quiet but consequential line between human expression and machine-generated imitation. Snapchat, YouTube, LinkedIn, and Substack are each, in their own way, choosing to deprioritize wholly synthetic content — what the industry now calls 'AI slop' — not by banning the tools that create it, but by refusing to amplify the flood it produces. The move reflects a deeper anxiety: that when fakeness is rewarded at scale, trust in everything erodes, and the shared sense of reality that makes communication meaningful begins to dissolve.
Four major platforms — Snapchat, YouTube, LinkedIn, and Substack — have launched a coordinated effort against 'AI slop': the industrial-scale flood of low-quality, entirely synthetic content that generative AI has made trivially easy to produce. Snapchat announced it will stop promoting wholly AI-generated videos in its Spotlight feed, joining the others in a shift that signals something larger than any single policy change.
The concern driving all four is the same: trust is eroding. Research cited by Substack's CEO Chris Best suggests that up to 40% of social media writing is now fake or AI-generated, and the more synthetic content people encounter, the less they believe anything online is real. 'Platforms that reward fakeness will create a race to the bottom,' Best said, unveiling a new tool to help readers identify AI-written text.
LinkedIn has been confronting the problem at enormous scale, blocking billions of AI-generated comment attempts in recent months and stopping hundreds of thousands of automated posts every day. This week it introduced a user-facing reporting button for suspected machine-generated content, while quietly removing a feature that had prompted users to 'enhance' their posts with AI — replacing it with a simpler proofreading tool.
YouTube's approach targets the money. After researchers found entire channels built on AI-generated content earning millions in ad revenue, the platform updated its policies to demonetize videos that are generic, repetitive, or template-based — precisely the formats generative AI produces fastest. Trust and safety chief Matt Halprin acknowledged the tension: the same technology enables genuinely creative work and industrial content farming alike.
What unites these moves is not a hard line against AI, but a subtler and perhaps more durable choice: these platforms are deciding their purpose is to surface human-made work, not to serve as distribution networks for synthetic material. By removing the friction that makes feed-flooding easy and giving users tools to identify and report it, they are collectively wagering that their audiences value authenticity over volume — and that the long-term health of the internet depends on that bet being right.
Four of the internet's largest platforms have begun a coordinated push against what the industry now calls "AI slop"—the flood of low-quality, entirely synthetic content that generative AI tools make possible to produce at scale. Snapchat announced on Friday that it would stop promoting wholly AI-generated videos in its Spotlight feed, joining YouTube, LinkedIn, and Substack in a series of moves that, taken together, signal a shift in how social media companies think about the content they amplify.
The problem is straightforward enough. As generative AI tools have become easier to use and more capable, they've made it trivial to produce essays, images, and videos at industrial volume. The result is a glut of repetitive, low-effort content flooding social feeds—material that research consistently shows users don't actually want to see. Snap's own assessment was blunt: entirely AI-generated content tends to be low-quality and repetitive, and it doesn't reflect what Snapchat's audience is looking for. The company isn't banning AI tools outright; Snapchat offers its own AI features for editing and enhancing content, and that material will continue to appear in recommendations. The distinction is crucial: AI-assisted content stays, but content that is wholly machine-generated gets deprioritized.
The broader concern animating all four platforms is erosion of trust. Research cited by Substack's co-founder and chief executive Chris Best suggests that up to 40 percent of writing on social media is now fake or AI-generated. More troubling still, the more synthetic content people encounter in their feeds, the less likely they become to believe that anything they're seeing online is genuine. "It's getting harder to tell what's real on the internet," Best said last week, announcing a new tool designed to help readers identify AI-written text. "Platforms that reward fakeness will create a race to the bottom."
LinkedIn, the professional networking platform, has been catching the problem at scale. The company's chief product officer reported that in recent months alone, the platform had blocked billions of attempts to post AI-generated comments. Every single day, LinkedIn is now stopping hundreds of thousands of automated comment attempts. This week, the platform introduced a reporting button that lets any user flag posts or comments that appear to be machine-generated. At the same time, LinkedIn is removing an automated prompt that used to offer users the option to "enhance" their posts with AI. That feature is being replaced with a simple proofreading tool—a deliberate step backward from automation.
YouTube's response has focused on monetization. Research from last year uncovered entire channels dedicated solely to AI-generated content, some with millions of subscribers and millions of dollars in revenue. Earlier this month, YouTube updated its policies to prevent three categories of videos from earning money: those that are generic, repetitive, or template-based. These are precisely the formats that generative AI makes easiest to produce at speed. Matt Halprin, YouTube's trust and safety chief, acknowledged the paradox: the same technology that enables creators to do genuinely useful work also enables what amounts to content farming. "The same technology really enables great stuff," he said. "But it also enables stuff that's kind of content farming, and that's the stuff that we don't want to have."
What's notable about this moment is not that any single platform has taken a hard line against AI. None of them have. Snap, LinkedIn, and YouTube all continue to allow AI-enhanced or AI-assisted content. The shift is more subtle and perhaps more durable: these platforms are collectively deciding that their role is to surface human-made work, not to become distribution networks for synthetic material. They're removing the friction that makes it easy to flood feeds with machine-generated posts. They're deprioritizing content that is entirely synthetic. They're giving users tools to identify and report it. In doing so, they're betting that their audiences value authenticity more than volume, and that the long-term health of their platforms depends on it.
Citações Notáveis
It's getting harder to tell what's real on the internet.— Chris Best, co-founder and CEO of Substack
The same technology really enables great stuff. But it also enables stuff that's kind of content farming, and that's the stuff that we don't want to have.— Matt Halprin, YouTube's trust and safety chief
Quer a matéria completa?
Leia o original em BBC News ↗
Technology