AI Usage
Scale
UK

Ця довідкова сторінка доступна англійською мовою.

Independent proposal · v1.3.1

Project status

A shared vocabulary for describing how generative AI contributed to a work. Definitions, tools and the evidence behind the project.

Project status

Scope and stewardship

Vocabulary
6 levels
Core languages
22
Definitions & marks
CC0

The scale provides six definitions, a decision procedure, downloadable marks and machine-readable data. English is the normative source. A declaration records a publisher's account of the process; it does not certify quality, rights or legal compliance.

Konrad Karauda maintains the project in the public repository. Decisions and contribution routes are documented in the governance and contribution guide.

Public references

Evidence checked on 10 October 2026. These records show references and integration interest; an independent production declaration has not yet been verified.

  • CHAOSS disclosureRelated tool

    The scale is listed among other AI disclosure and attribution tools.

  • CHAOSS AI AlignmentResource list

    A second reference from the same community, in its moderation resources.

  • diarieDeferred integration

    The public backlog records adoption interest and a question about the Level 3 / Level 4 boundary.

To record a use, provide a public declaration in the adoption evidence file. Using the scale requires no signature, registration or account.

Standards discussion

Schema.org's DefinedTermSet and DefinedTerm describe the vocabulary. Work declarations use an explicit project extension for aiUsageLevel.

The relevant public venues are Schema.org's AI disclosure discussion and the W3C AI Content Disclosure Community Group. Follow the submission and pilot record for requests, responses and implementation evidence.

The scale remains an independent proposal. Suggested mappings do not imply institutional acceptance. Google lists its supported features in the Search gallery; a dedicated feature for these levels is not documented there.

Review and next steps

Editorial review
The October revision was developed with AI. Human editorial review is pending.
Translation review
The 21 translations are labelled as machine translations awaiting fluent human review.
Classification boundary
The Level 3 definition requires meaningful review, but its current decision path stops before asking about it. An automated Level 3 result is provisional.

The next priorities are independent pilots, fluent language reviewers and a documented decision on the review boundary. See the roadmap. The original July rationale is retained in the historical archive.