A Regular Post Changes AI Search Answers

Researchers (arXiv 2610.11932) analyzed a sample of 17,211 citations from 10 AI search platforms and showed that citations concentrate in a narrow circle of easily accessible sources, and that regular posts on such platforms change search results — up to the point of citing a deliberately fabricated concept within 7 days.

Compiled by Sergey KostenchukPublished 2026-10-10Updated 2026-10-10

2026-10-10 Research

🤖 A Regular Post Changes AI Search Answers

In the preprint arXiv 2610.11932 from October 8, researchers analyzed 17,211 citations across 10 AI search platforms covering 6,356 domains: the top-20 "own" domains account for 20.5–70.8% of citations, 15 of 22 platforms have low publication barriers, and 8 of 10 platforms in experiments cited a deliberately fabricated concept within 7 days.

🌍 Posts on easily accessible UGC platforms have become a content injection channel: one post on a "preferred" platform yielded more citations than 20+ posts on the rest, and buying a GEO service (Generative Engine Optimization) for $14 yielded a citation with control markers within an hour.

👤 "Sources" in an AI search answer may be seeded or paid posts — open the original source and check who and when published it. Caveat: the GEO case is a single example.

Source 1: https://arxiv.org/abs/2610.11932 Source 2: https://arxiv.org/html/2610.11932v1