Snapchat, YouTube, and LinkedIn have joined a growing movement, including Substack, to tackle the spread of fake texts, images, and videos created entirely by artificial intelligence tools. This material, known in the industry as "AI slop" or "AI junk," has rapidly proliferated due to the easy access to programs that can generate everything from complete articles to hyper-realistic videos in seconds.
Snap, the owner of Snapchat, has decided to stop promoting videos generated exclusively by AI in its popular Spotlight feature, now prioritizing "authentic, human-produced content." The company clarified that it does not prohibit the use of the technology; materials that use AI only as an aid in editing or enhancement will remain visible in the feed. However, Snap admitted that fully automated content is often repetitive, low quality, and misaligned with what the community wants to consume.
Recent research on user behavior, as reported by the BBC, confirms this dissatisfaction. An additional study revealed a worrying consequence: the greater the exposure to fake AI content on social media, the lower users' trust in the veracity of any other online publication becomes.
LinkedIn and Substack's Stance
This scenario reflects a widespread discomfort. Chris Best, CEO and co-founder of the newsletter platform Substack, expressed his frustration by presenting a tool for readers to identify algorithmically created texts, stating: "It is becoming increasingly difficult to know what is real on the internet."
Best cited data indicating that up to 40% of texts published on social media today are false or automated by AI. The executive warned that "platforms that reward fake content will create a race to the bottom."
On LinkedIn, which focuses on the corporate world, the response was equally vigorous. The network implemented a feature that allows users to report posts or comments suspected of being made by bots. According to Hari Srinivasan, LinkedIn's product director, combating "AI junk" has become a top priority. In recent months, the platform has blocked billions of automated comment attempts and intercepted hundreds of thousands of spam daily. Additionally, LinkedIn removed an automatic shortcut that suggested users "improve text with AI" before posting, reverting to offering only a basic grammar checking tool.
Changes on YouTube
YouTube has also adjusted its monetization rules to discourage the filling of the platform with what it calls "inauthentic content." Investigations conducted over the past year exposed the existence of hundreds of channels operated exclusively by AI, many of which reached millions of subscribers and generated significant profits.
To control the growth of these "content farms," the Google-owned video platform defined three criteria for videos that will lose the right to receive advertising revenue: productions that are generic, repetitive, or based on standardized templates. Matt Halprin, head of YouTube's Trust and Safety division, emphasized in an interview that while artificial intelligence can be a valuable creative ally, it is being used abusively for the massive and low-quality production of videos. He concluded that "The same technology that allows creating amazing things also enables the creation of disposable material aimed only at 'mining' views. And it is exactly this type of content that we do not want on our network."