Snapchat joins other popular platforms in fight against 'AI slop'
BBC ● Covered by 2 sources
Snapchat's cutting wholly AI-made videos from its Spotlight recommendations, favoring human content instead. It joins YouTube, LinkedIn and Substack in a bigger pushback against fake, low-quality 'AI slop' flooding feeds.
Based on reporting by BBC — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Snap dropped a small but telling policy change on Friday: its Spotlight feature, the algorithmic feed that surfaces short videos to Snapchat's users, will no longer recommend content that's entirely AI-generated. Human-made stuff gets the priority now. Videos that were merely touched up with Snap's own AI editing tools are still fair game, but fully synthetic clips are getting pushed to the back of the line.
This isn't happening in a vacuum. Substack, LinkedIn and YouTube have all rolled out their own countermeasures in just the past two weeks, which suggests something has shifted industry-wide rather than at one company. Substack's CEO Chris Best launched a detection tool for readers and pointed to research claiming up to 40% of social media writing is now AI-generated or fake. His line was blunt: platforms that reward fakeness are running a race to the bottom.
LinkedIn's approach has been more aggressive on the enforcement side. The company says it has blocked billions of AI-generated comment attempts over the last couple of months alone, and is now catching hundreds of thousands of automated posting attempts daily. It also added a simple report button for flagging suspected AI content, and killed off the one-click 'enhance with AI' prompt that used to appear when users drafted posts, reverting to a basic proofreading feature instead.
YouTube's version of the crackdown targets money rather than visibility. Earlier this month it named three categories of video, generic, repetitive, and template-based, that will no longer qualify for ad revenue. That matters because researchers found channels built entirely from AI-generated content pulling in millions of subscribers and, in some cases, millions of dollars. YouTube's trust and safety lead, Matt Halprin, framed it as separating genuine creative use of AI tools from what he called content farming, the kind of high-volume, low-effort output that's cheap to produce at scale and increasingly clogs up recommendation feeds.
What ties all four moves together is a shared bet that authenticity, not sheer volume of content, is what keeps people opening the app. Surveys cited in the reporting back that up: the more synthetic content people spot in their feeds, the less they trust anything they see there, real or not. That's a slow-building problem for platforms whose entire business model depends on people believing what's in front of them.
My take — AI-written commentary, not fact-checked reporting
Credit where it's due, but none of these four companies are doing this out of civic duty, they're doing it because trust is the product and slop was eating the product alive. The real test isn't the policy announcement, it's whether these platforms keep enforcing once the novelty wears off and the AI tools get good enough to dodge detection. Expect a permanent cat-and-mouse game, not a solved problem.
Read more about this at: BBC