About

A public evidence record built for reliable reuse.

AI Experiment Log publishes structured records of Experiments in which AI contributed to planning, execution, analysis, or documentation, so that the work can be checked and reused later. When humans perform physical operations or observations, or when automation performs part of the workflow, those roles are recorded separately from AI processing. The primary audience is later AI-assisted research and investigation, while human readers are equally within scope.

Each record distinguishes what was observed or verified, what is inferred, what remains unresolved, and what was not performed or was outside scope. The intention is not to rush toward broad claims, but to preserve enough context for a later investigator to judge the conditions and limits under which a result can be reused.

The preservation model, Evidence State, Experiment Status, and public-page structure are described in Record & publication method. The framing of published information, reuse, commercial independence, and information withheld for privacy or third-party rights are described in Disclosure & Privacy.

English is the default record language. Japanese pages are corresponding localisations with their own URLs and an explicit language relationship.