Snapchat joins other popular platforms in fight against 'AI slop'

Snapchat has joined other major platforms in the effort to combat "AI slop," the term for AI-generated content. The company announced it will no longer recommend entirely AI-generated videos in its popular Spotlight feed, prioritizing "authentic, human-made content" instead. This move follows similar initiatives by YouTube, LinkedIn, and Substack in the past two weeks.

The proliferation of AI-generated content, ranging from essays to realistic videos, has surged as the tech industry rapidly develops user-friendly generative AI tools. Snap, Snapchat's parent company, stated on Friday that its platform will still recommend AI-enhanced or edited content, as it offers its own AI tools for content modification. However, Snap acknowledged that fully AI-generated content is often "low-quality," "repetitive," and generally not what users desire.

Recent studies support this view, indicating that people tend to agree with these descriptions of AI-generated content. Furthermore, a separate survey found that increased exposure to fake AI-generated content on social media feeds diminishes users' perception of the genuineness of any online content they encounter.

Chris Best, co-founder and CEO of Substack, highlighted the growing difficulty in discerning real content online. Last week, while introducing a new tool to help readers detect AI-generated writing on the newsletter platform, Best cited research suggesting that up to 40% of social media writing is now fake or AI-generated. He warned that "Platforms that reward fakeness will create a race to the bottom."

LinkedIn, a professional social media platform, this week implemented a feature allowing users to report posts or comments that appear to be AI-generated. Hari Srinivasan, LinkedIn's chief product officer, stated that "AI slop is a top priority for all of us." He revealed that the platform has "blocked billions" of attempts to post AI-generated comments in recent months, now catching "hundreds of thousands of automated comment attempts" daily. Like Snap, LinkedIn is not entirely rejecting AI use; users who employ AI tools to "refine" their posts will not be targeted. However, LinkedIn is removing an automated prompt that offered to "enhance" posts with AI, reverting to a simple proofreading tool.

YouTube, Google's video platform, also updated its policies this month regarding content monetization for creators. Last year, research identified numerous YouTube channels consisting solely of AI-generated content, many with millions of subscribers and significant revenue. To prevent the platform from being inundated with "inauthentic content," YouTube now categorizes videos that cannot be monetized into three types: generic, repetitive, or template-based. These styles are easily and quickly produced using generative AI tools.

Matt Halprin, YouTube's trust and safety chief, explained in an interview about the updated rules that while AI tools can be beneficial, they are more frequently used to create fake and low-quality videos. "The same technology really enables great stuff," Halprin said, "But it also enables stuff that's kind of content farming, and that's the stuff that we don't want to have."