Major social platforms are tightening their content rules in response to a wave of machine-generated posts, with Snapchat becoming the latest to filter what the BBC describes as "AI slop" from its most visible feed. The move reflects a growing acknowledgement among platform operators that entirely AI-created material is pushing users to question whether anything they see online is genuine.
Why platforms are cracking down
According to the BBC, Snap, the parent company of Snapchat, said on Friday it would stop recommending "wholly AI-generated videos" in its Spotlight feed in favour of "authentic, human-made content." Snap did not prohibit all AI output — content "enhanced or edited" with AI tools will still appear in recommendations — but the company acknowledged that entirely AI-generated video is typically "low-quality" and "repetitive", and is not what users want.
Over the past two weeks, YouTube, LinkedIn and Substack unveiled similar strategies, the BBC reported. The simultaneous moves target content that has proliferated as generative AI tools make it easy to produce essays, images and videos at scale.
What each platform is doing
| Platform | Action | Details per BBC report |
|---|---|---|
| Snapchat | Spotlight recommendation change | Stop recommending wholly AI-generated videos; AI-enhanced or edited content remains eligible |
| YouTube | Monetisation policy update | Videos that are generic, repetitive or template-based will not be allowed to make money |
| AI-reporting button and comment blocking | Users can flag posts as AI-generated; billions of automated AI comment attempts blocked in recent months | |
| Substack | AI-writing detection tool | Readers can detect AI-generated writing; CEO cites research that up to 40% of social media writing is fake or AI-generated |
The credibility problem
The BBC reported research showing that the more AI-generated content people see in a social media feed, the less likely they are to believe any of the content they are shown is genuine. A separate survey found audiences broadly agree with descriptions of AI slop as low-quality and repetitive.
"It's getting harder to tell what's real on the internet."
Chris Best, co-founder and chief executive of Substack, said that last week while announcing the new tool, adding: "Platforms that reward fakeness will create a race to the bottom."
Platform statements
LinkedIn's chief product officer, Hari Srinivasan, wrote on the platform that "AI slop is a top priority for all of us." In just the last couple of months, the BBC reported, LinkedIn had "blocked billions" of attempts to post AI-generated comments. "Every day we are now catching hundreds of thousands of automated comment attempts," Srinivasan said.
LinkedIn is not rejecting AI entirely: users that employ AI tools to "refine" their posts should not be caught up in the anti-slop effort. However, LinkedIn is removing the automated prompt that offered to "enhance" posts through AI, returning to a simple proofreading tool.
YouTube, owned by Google, updated its monetisation policies this month, according to the BBC. Research from last year found scores of channels solely running AI-generated content, many with millions of subscribers and some making millions of dollars in revenue. Matt Halprin, YouTube's trust and safety chief, said in an interview that AI tools can help people with their content, according to the BBC. The updated rules split ineligible videos into three categories: generic, repetitive or template-based.
What this means for organisations
For enterprises that distribute content through social platforms or monitor those networks for brand and reputation risk, the BBC's reporting signals that platform owners now treat high-volume AI-generated material as a trust liability. The survey cited by the BBC found that exposure to fake content makes users doubt the authenticity of everything else they see — a direct threat to any organisation using these channels to communicate with customers or partners. The new moderation tools and policies from Snapchat, YouTube, LinkedIn and Substack amount, as Srinivasan put it, to a top-priority cleanup of machine-generated noise.
Sources:
