Ordinary web posts can enter AI-search answers
In plain words
This paper presents a way to measure whether ordinary posts on public websites can become sources in consumer AI-search answers. The researchers checked which websites AI search cited, how easy it was to publish on related platforms, and whether marked posts later appeared in answers.
In a test about one made-up concept, posts on selected platforms were cited by eight of ten AI-search systems within seven days. That result shows a route under those test conditions; it does not show that the posts were true or that the same result holds for other topics.
The research
When an AI-search answer cites a page, readers may treat that source as evidence. It remains unclear whether a repeatedly cited domain can also be easy for ordinary publishers to use, letting publication access feed into the sources AI search selects. Repeated citation alone does not establish accuracy or trustworthiness.
The framework maps cited domains, tests publishing barriers and tracks marked posts to measure when ordinary publication enters AI-search answers.
The authors mapped citations from 120 designed queries across ten AI-search platforms and tested 22 publishing platforms linked to cited domains. Citations concentrated in platform-specific sources, while many tested platforms had low or medium setup barriers and quick or no visible review.
On selected citation-preferred platforms, posts about one fabricated concept were cited by eight of ten systems within seven days. This demonstrates a route into answers under those conditions, not that the pages were accurate.
These live-web measurements cover a designed query panel, selected topics and one March–April 2026 campaign. The results do not establish a general rate or a rule for how many posts on one platform equal a post on another, and platform behavior may change.