Social platforms tighten rules on AI-generated content

Social platforms are trying to balance their investment in AI tools with the need to keep feeds from being overrun by low-quality machine-made posts. Companies that once competed for creators with monetization programs are now clarifying rules for AI-generated content, particularly where creators can earn money.

YouTube said on July 20 that it wants to reward creators who use AI to enhance original storytelling, not those mass-producing generic videos or AI personas. A spokesperson said the move clarifies its existing inauthentic content policy after creators were demonetized for repetitive, low-effort, or emotionally manipulative videos. The platform has removed channels posting AI-generated videos, and Kapwing estimated that 21% of YouTube’s feed was AI-generated videos.

TikTok is offering AI tools through its app and CapCut while requiring labels for AI-generated content, adding invisible watermarks to some videos, and testing better detection of accounts posting AI spam. Substack is adding AI-use estimates through Pangram, though publishers can opt out. Pinterest labels suspected AI content and gives users controls to see less of it, while Meta applies “AI info” labels across Facebook, Instagram, and Threads and has expanded labels to AI-powered ads.