# Selekoa > A documentary history of machine intelligence. Selekoa is a machine-first starting point for research on the history of machine intelligence: machines find and check the records, and every one links to its source, for you to verify and build on. It records people, institutions, documents, funding and governance as individual claims. A claim is one scoped statement linked to the source passages it rests on, with its date, qualification and review state. The site’s pages are drawn in the browser from hash addresses such as `#/about`, so a program that does not run JavaScript sees only the start page. The records behind every page are published as the files below. ## Before relying on a record - A claim records what its sources state. It is not a finding of fact. Read the cited passage in the original source, in the edition the evidence names. - The records are machine-checked: an automated research process checked each passage. Most claims have not been reviewed by an editor (`editorialState: seed`); the README lists what the labels mean. - Coverage is selective and uneven. A missing record is a gap in coverage, not evidence that something did not happen. - A published framework, commitment or law shows what its author said, not that anyone implemented or complied with it. - Selekoa’s own records and text are licensed under CC BY 4.0. Quoted source passages (the `quote` field of evidence records), quotations inside Selekoa text and the source documents are excluded and remain subject to their owners’ rights. The [scope note](https://selekoa.org/data/LICENSING.md) and the [legal code](https://selekoa.org/data/LICENSE-DATA.txt) are published beside the data. ## Data: dataset release v0.7.2 (2026-10-26) These are the files of the latest release, unchanged. The JSON files are authoritative. These addresses always serve the latest release; `manifest.json` names its version. - [README.md](https://selekoa.org/data/README.md): what the release contains, how to cite a record and what to check before relying on one - [manifest.json](https://selekoa.org/data/manifest.json): version, release date, counts, schema version, each file’s SHA-256 checksum and each field’s text class - [claims.json](https://selekoa.org/data/claims.json): one scoped statement each, with dates, status and evidence IDs - [entities.json](https://selekoa.org/data/entities.json): people, organisations, documents, concepts, models, public bodies and other referents - [evidence.json](https://selekoa.org/data/evidence.json): one source passage used for one claim, with its locator, quote or paraphrase, source role and check date - [sources.json](https://selekoa.org/data/sources.json): the inspected bibliographic expression of each source - [collections.json](https://selekoa.org/data/collections.json): authored reading collections and the claims each one uses - [sources.csl.json](https://selekoa.org/data/sources.csl.json): every source as a CSL-JSON item, to cite the original documents - [Dataset archive (ZIP)](https://selekoa.org/release/selekoa-dataset-v0.7.2.zip): every file of this release with the licence files. Each release has its own archive at a fixed address of the form `release/selekoa-dataset-vX.Y.Z.zip`. ## How to cite and contact - [How to cite](https://selekoa.org/data/README.md): cite the original source for a fact, and Selekoa for a record, by its record ID and release version: “Selekoa, claim record ``, dataset release v0.7.2 (2026-10-26).” - [About & method](https://selekoa.org/#/about): method, labels, known limits, citation and corrections (needs JavaScript) - Corrections and requests: maintainer@selekoa.org. Give the record ID, what is wrong and the source that shows it. ## Optional - [changes.json](https://selekoa.org/data/changes.json): records added, changed and removed since the previous release, by ID - [CHANGELOG.md](https://selekoa.org/data/CHANGELOG.md): the same changes in prose - [CITATION.cff](https://selekoa.org/data/CITATION.cff): citation metadata for the release as a whole - [claims.csv](https://selekoa.org/data/claims.csv): flat view of `claims.json` - [collections.csv](https://selekoa.org/data/collections.csv): flat view of `collections.json` - [entities.csv](https://selekoa.org/data/entities.csv): flat view of `entities.json` - [evidence.csv](https://selekoa.org/data/evidence.csv): flat view of `evidence.json` - [sources.csv](https://selekoa.org/data/sources.csv): flat view of `sources.json`