Dianping removed 11.6M AI reviews in a year: platforms fight AI slop as AEO spam games LLM answers
创业邦 · wechat · 2026-10-02
AI-generated content is diluting platform value: Dianping processed 11.6M AI-written reviews last year, Deezer says nearly half of new daily song uploads are AI-generated, and Spotify removed 75M spam tracks. LinkedIn's new "looks like AI slop" report button cut AI career-article views by 40%, while Snapchat, Pinterest, TikTok and Instagram are all limiting pure AI content in recommendations.
A sneakier trend is AEO/GEO replacing SEO: marketers plant detailed fake reviews in UGC communities like Reddit because that's exactly what LLMs cite. Cornell Tech research found 20% of AI research tools' sources come from UGC (Reddit alone 12% of crawled URLs), and injecting a "doctored copy" of a high-frequency indexed page puts planted content into final AI answers 38-51% of the time (up to 62% with multiple sources) — the shortest payload was just 13 English words.
The author argues AI paradoxically creates a "human premium": producers of generic content get cheaper, while people with real experience, judgment and taste become more valuable.
More from AGI Musings
- Terence Tao's Caltech lecture slide sparks debate: do cancer drug analogies sink the interpretability argument? — ctjlewis · 2026-10-11
- Spanish AI founder publishes 'Vita', an essay on AI, science and the limits of life — uribeetxebarria · 2026-10-11
- AI slop is a leadership failure first: studies of 10,000 leaders point to missing standards — DrKavner · 2026-10-11
- AI governance scholar warns the superintelligence race may be only ~8 months from its target — LuizaJarovsky · 2026-10-11
- We trust interpreters over raw information — and AI could help us trace claims instead — CyborgWriter · 2026-10-11
- Musk envisions a 'sentient sun': harnessing solar energy to build digital superintelligence — XFreeze · 2026-10-11