EST. MMXXVI · THE INSTITUTION OF RECORD FOR AI CITATION · OPEN METHODOLOGY
THE AI CITATION INSTITUTEEx corpore, auctoritas.the record
Most of this category asserts. We measure, and then we publish the measurements — methodology, numbers, and the honest limitations. Every record is accession-numbered, permanent, and open access. Quote it, chart it, cite it; all we ask is attribution.
The register a page is written in changes how often an AI cites it. Across 618 trials, a casual-precise voice beat a corporate one — a +33 percentage-point lead on Claude. Sloppy numbers reproduced 0% of the time; addressing the AI read as distrust.
read the recordA 91-post, four-engine audit sorted by inbound internal links: pages with none were cited 4.5% of the time; pages with ten or more, about 44% — a tenfold swing on the same content. Perplexity had the widest reach at about 59%.
read the recordA ~4,000-trial study across Claude, GPT and Gemini. Rigor markers on your numbers are the strongest on-document lever; independent corroboration is the bigger off-document one; and an undisclosed conflict of interest collapses top-two inclusion from 99% to 4%, with a disclosed editorial firewall recovering most of it in the language tests.
read the recordThe research feeds the live scoreboard and the answer library. Next candidates: schema-type × citation correlation, and a new-domain time-to-first-citation longitudinal — with aicitationinstitute.org itself as the subject, logged from day zero.