- The vendors’ own customers show it. 36% of brands in one popular AI-content tool’s success-story section show the “Mount AI” traffic pattern in Google (sharp rise then massive drop); at another platform it was 3 of 4 featured brands (75%).
- Grokipedia is the marquee case: it lost Google visibility in late January/early February 2026, and answer engines (ChatGPT, AI Mode, AI Overviews) reduced Grokipedia citations at exactly the same time — see Google rankings affect AI visibility.
- Google can detect low-effort AI content. Google’s own 2020–2021 research tested AI detectors on 500 million pages; a researcher from that study now works on AI-content detection at DeepMind. Google’s 2025 Quality Rater Guidelines explicitly mark mass-produced, no-original-value pages “Lowest” quality, and Originality.AI found 100% of sites removed in Google’s 2024 spam update carried AI-generated posts.
- User signals compound it: bounces from slop pages feed back into rankings (confirmed for Bing, and for Google via patents and antitrust revelations).
The real risk of low-quality AI-generated content
Peec’s analysis (Feb 2026) — sites publishing raw AI content at scale suffer Google visibility drops, including 36–75% of AI-content tools’ own success stories.
Publishing raw AI-generated content at scale can tank online visibility. Peec’s analysis (full article, Feb 2026) documents the pattern: sites with poor-quality AI content rank for a while, then drop — sometimes dramatically.