For the past few years, social platforms have spent a lot of time figuring out how to help people create content with AI. And lately, it seems as though they’re figuring out what to do with all of that AI-generated content once it starts filling people’s feeds.
LinkedIn recently added a new option that lets users report posts that “seem like AI slop,” referring to low-quality, mass-produced AI content. But it’s far from the only platform making changes.
Substack has added AI detection for written content. Snapchat is prioritizing original content over synthetic AI videos. YouTube has started automatically detecting and labeling some AI-generated videos.
These updates signal a broader shift in how platforms handle AI-generated content — decidedly not by banning it but by giving users more information and control over what they see. Let’s talk about it.
LinkedIn Is Letting Users Report “AI Slop”
LinkedIn is one of the latest platforms to address the issue.
The company recently added a “Seems like AI slop” option to its reporting menu, allowing users to flag content they believe is low-quality and AI-generated. That’s among a few other changes the platform is making to reduce low-quality AI content.
It’s worth making an important distinction here: AI-generated content and “AI slop” aren’t necessarily the same thing. Someone might use AI to research a topic, edit their writing or help create an image without producing what most people would consider “slop.” The latter term is generally used to describe repetitive, low-effort and/or mass-produced AI content created primarily to attract attention or engagement.
LinkedIn’s new reporting option gives users a way to make that distinction themselves or, at least, tell the platform when they think a post crosses that line. And LinkedIn isn’t alone in trying to draw it.
Substack Is Putting AI Detection in Readers’ Hands
Substack recently took a different approach.
In July, the publishing platform announced an integration with AI detection company Pangram. The feature allows users to scan posts, comments and replies to get an estimate of how much was written by a person versus AI.
That makes Substack’s approach somewhat different from the AI labels we’ve gotten used to seeing elsewhere.
Instead of relying exclusively on creators to disclose AI use or the platform to automatically attach a label, Substack is giving readers an AI detection tool they can use themselves.
Of course, as many marketers and non-marketers know, AI detection isn’t perfect and results should never be taken as certain. Still, the feature represents another way platforms are trying to make AI’s role in content more visible to audiences.
Snapchat Is Prioritizing Original Content
Snapchat is taking yet another approach: changing what people see in the first place.
Earlier this year, the picture messaging platform said it was adjusting Spotlight, its short-form video feed, to put greater emphasis on original content created using the Snapchat camera.
As part of that change, users should see fewer synthetic AI videos and widely syndicated posts. Snap said the goal is to emphasize “fresh, personal content that feels native to our community.”
The company has since gone further. In late July, Snap announced changes to its Spotlight rewards program aimed at rewarding what it describes as “authentic creativity” as low-quality, repetitive and AI-generated content becomes more common online.
Unlike LinkedIn, this isn’t primarily about asking users to identify AI content. And unlike Substack, Snap isn’t giving people a detector. It’s using recommendation and creator incentives to put greater emphasis on original content.
YouTube Is Automatically Detecting Some AI Content
YouTube, meanwhile, has been expanding its AI labeling system.
Creators were already required to disclose realistic altered or synthetic content in certain circumstances. But in May, YouTube began rolling out internal signals designed to identify AI-generated content automatically.
If a creator doesn’t disclose AI use but YouTube detects significant photorealistic AI-generated content, the platform can now automatically apply a label.
YouTube also made those labels more prominent, placing them directly beneath long-form videos and overlaid on Shorts, which is a notable change from other systems that depend primarily on creators to report their AI use.
It also demonstrates just how many different parts of the content ecosystem platforms are now looking at, as detection systems, recommendation algorithms and user reports become increasingly involved in AI transparency.
Other Platforms Are Giving Users More Control Over AI
All of these examples don’t appear to be isolated experiments, either.
Pinterest, for example, introduced labels for AI-generated or modified images in 2025. Its systems use both metadata and visual classifiers to determine when a Pin should receive an “AI modified” label.
The platform later introduced controls allowing users to reduce the amount of generative AI content appearing in categories including art, beauty, fashion and home decor. TikTok has also been developing ways to identify and label AI-generated content.
And the pressure isn’t coming exclusively from users or platforms themselves.
As of August 2, 2026, certain transparency requirements under the European Union’s AI Act have taken effect. Among other provisions, the rules establish disclosure requirements around certain AI-generated and manipulated content.
So, while each platform has its own reasons for making these changes, AI content transparency is increasingly becoming part of the wider regulatory environment as well.
There’s No Single Approach to AI-Generated Content
What’s interesting about these changes isn’t just how different they are in their approach, but also that none of them (yet) have amounted to a blanket rejection of generative AI. After all, that would be a bit counterintuitive since many of these platforms offer AI-powered creation features of their own.
Instead, platforms appear to be developing new layers of infrastructure around AI-generated content: identifying it, labeling it, allowing users to report it and, in some cases, changing how frequently it gets recommended.
That also introduces some new challenges. AI detection systems alone can and often do make mistakes, which might help explain why some platforms are combining automated detection with creator disclosures, appeals or other mechanisms.
What Does This Mean for Marketers?
For marketers, the immediate takeaway is simple: how we use AI in content creation is becoming more visible to audiences, but that shouldn’t diminish its value in the process … so long as you’re not contributing to the slop problem, i.e., low-effort, repetitive content.
AI can support the creation of original, authoritative content, playing strong roles in research, ideation, structuring, editing, analysis and even writing.
Content that is clearly original, well-researched and shaped by human perspective is less likely to be deprioritised, even if AI tools were used along the way.
Final Thoughts
Despite these platforms’ actions, AI-generated content won’t disappear from social media anytime soon.
But they now have to figure out how that content fits alongside everything else.
So far, there doesn’t seem to be a universal solution. Some are adding labels. Others are building detectors, adjusting recommendation algorithms or giving users more control.
LinkedIn’s “AI slop” button might be the bluntest example yet, but it’s part of a much broader pattern happening across the internet: figuring out how to distinguish AI-generated content, how prominently to surface that information and, increasingly, whether users want to see that content at all.
