Tang, J., Yu, X., Wikström, P., & Pappalardo, K. (2026). Behind the “AI slop”: exploring the AI-generated viral short video production in China. AI & SOCIETY. https://doi.org/10.1007/s00146-026-03254-9
Summary
This paper examines “AI slop” creators in China who deploy generative AI tools to systematically manufacture viral short-form video content. Rather than treating virality as an unpredictable stroke of fortune, these creators reframe it as a manageable, engineered outcome. The authors argue that this represents an emerging form of platform-based creative labour, one that destabilises conventional understandings of creativity, authorship, and what it means to be a “creator.” Using the Chinese short-video ecosystem as a case study, the paper documents how AI-driven production has become industrialised and formalised into a distinct mode of cultural production.
Key Contributions
- Introduces and characterises “AI slop” creators as a distinct category of creative labourers within China’s platform economy.
- Advances scholarly understanding of how generative AI reshapes creativity, authorship, and the concept of the “creator.”
- Provides an empirical account of how virality is systematically managed and engineered within AI-driven cultural production.
Methods
The study draws on ethnographic research focused on AI slop creators in China. It analyses concrete creator practices, including benchmarking viral content, testing content formats, operating “matrix” account systems for coordinated distribution, and formalising production workflows into repeatable processes.
Findings
- Creators use generative AI tools to reverse-engineer and replicate viral formats at scale.
- Matrix account systems enable coordinated, high-volume content production and distribution across many accounts.
- Production processes are formalised and systematised, turning virality into an engineered rather than serendipitous result.
Connections
This paper contributes to the study of synthetic media by documenting the labour and infrastructure behind mass-produced AI content, connecting to broader concerns about how such content circulates and how audiences respond to it, as in Baym2026-tr and Manovich2026-ih. Its focus on the reshaping of creative work and authorship also speaks to debates on the labour impacts of generative AI.
Podcast
A research-radio episode discusses this paper: 🎧 MP3 · Spotify