ARTFEED — Contemporary Art Intelligence

Study Reveals Scale-over-Preference Dynamic in AI-Generated Content on Video Platforms

ai-technology · 2026-05-14

A study published on arXiv (2604.01690) analyzing a Chinese video-sharing platform with tens of millions of users found that AI-generated content (AIGC) creators achieve aggregate engagement comparable to human-generated content (HGC) creators through high-volume production, despite consumers preferring HGC. The research identified a 'scale-over-preference' dynamic and highlighted the role of algorithmic content distribution in moderating competing interests. The findings advocate for AIGC-sensitive distribution algorithms and precise governance.

Key facts

  • Study uses longitudinal dataset from a leading Chinese video-sharing platform
  • Dataset includes tens of millions of users
  • AIGC creators achieve comparable engagement through high volume
  • Consumers show marked preference for HGC
  • Algorithmic distribution moderates competing interests regarding AIGC
  • Study advocates for AIGC-sensitive distribution algorithms
  • Research published on arXiv with identifier 2604.01690
  • Paper title: 'Scale over Preference: The Impact of AI-Generated Content on Online Content Ecology'

Entities

Institutions

  • arXiv

Locations

  • China

Sources