AI usage notice.

1 · Scope of this notice

This notice covers the content published on extendedmatrix.org (this website) and on the documentation subdomain docs.extendedmatrix.org, including the manuals of Extended Matrix (the formal language), EM Tools, 3D Survey Collection, Heriverse and related projects hosted on Read the Docs.

It does not cover peer-reviewed articles and conference papers published under third-party editorial responsibility (journals, proceedings): those follow the AI disclosure policy of the respective publisher, which we honour at submission time.

2 · How AI tools are used in authoring

Large-language-model assistants (currently Anthropic's Claude, via the Claude Code and Claude Cowork interfaces) are used as an assistive drafting and editing tool in the authoring workflow of this site and of the manuals. Typical uses include drafting prose from bullet-point outlines, rephrasing existing text, cross-checking terminology across manuals, generating tabular summaries from source material, and helping enforce editorial constraints (length caps, heading structure, terminological consistency).

AI is not used to publish content directly. Every text edit, whether produced ex novo or refined by the assistant, is reviewed by a human author before it is committed to the repository and pushed to the site. The commit history of the source repositories is the audit trail of that review.

3 · Human review and editorial responsibility

Editorial responsibility for the content published on this site and on the Extended Matrix documentation is held by Emanuel Demetrescu, in personal research capacity, with the collaboration of the Extended Matrix contributors listed on the team page. Every publication decision — what to write, what to remove, what to endorse as the project's position — is made by a human author with subject-matter competence in the domains covered (Extended Matrix, stratigraphy, virtual reconstruction, 3D documentation, digital cultural heritage).

This satisfies the two cumulative conditions of the human-review-and-editorial-responsibility exception provided for in Article 50(4) of the AI Act: an informed human deliberately examines the substance of the content before publication, and a human person holds the ultimate legal responsibility over its publication.

4 · What this project does not do

• We do not deploy an AI chatbot or conversational agent that interacts directly with visitors: there is no LLM behind the scenes answering questions on the site. Any such tool, if introduced in the future, will be disclosed explicitly at the point of interaction, as required by Art. 50(1) of the AI Act.
• We do not publish AI-generated deepfake imagery, synthetic audio, or synthetic video, and we do not manipulate photographs of real people to represent events that did not occur (Art. 50(4), deepfake obligation).
• We do not use AI to generate content on matters of public interest (political, health, safety) outside of the project's technical and scientific scope, and we do not publish AI-drafted text without human review.

5 · Illustrations, figures and screenshots

The vast majority of figures on this site and in the manuals are screenshots of real software (EM Tools in Blender, EMStudio, Heriverse, yEd), diagrams authored by the contributors, or photographs of real archaeological contexts. Where an illustration has been generated or substantially altered by an AI image tool, the caption says so explicitly.

6 · Code and software artefacts

Some source code in the projects linked from this site (EM Tools, EMStudio, 3D Survey Collection, s3Dgraphy, Heriverse) is written with AI-assisted coding tools. As with prose, every change is reviewed by a human maintainer before it is merged and released. Where relevant, individual commits or pull requests carry a co-authorship trailer identifying the assistant. The AI Act's transparency obligations concern published content, not the internal development workflow of software; this section is included for completeness, not because it is required by Article 50.

7 · Corrections and feedback

If you spot a factual error, a mis-attribution, or a passage that reads as AI-generated in a way that misrepresents the project's position, please write to emanuel.demetrescu@cnr.it or open an issue on the site repository. The commit history of that repository is the public record of every editorial decision made on this site.

8 · Regulatory references

• Regulation (EU) 2024/1689 (AI Act) — full text on EUR-Lex.
• AI Act, Article 50 — transparency obligations for providers and deployers of certain AI systems.
• European Commission — Guidelines on transparency obligations (the interpretive guidance the Commission published in 2026).

9 · Changes to this notice

This notice may be updated when the authoring workflow changes (new tools introduced, new categories of content added, revised Commission guidance). The "last reviewed" date at the top of this page reflects the latest substantive update. The full revision history is available in the site repository.