Editorial transparency
AI Usage and Content Disclosure
Where artificial intelligence is used in producing this site, where it is not, and who is responsible for what you read.
- Version
- 1.0
- Last updated
- Language
- English (binding version)
- Governing law
- Romanian law and directly applicable EU law
Legal documents
1. Our position in one paragraph
We use artificial intelligence as a writing and research assistant, never as a publisher. No text, image or data reaches this website automatically. A named person at LAVALI ONLINE SHOP SRL reads, checks, corrects and explicitly publishes every entry, and this company holds editorial responsibility for everything you read here. Where AI helped produce something, we say so, because we would rather tell you than have you wonder.
2. The standard notice
Where AI assistance was used in producing an entry, that entry carries this notice:
AI-assisted, human-verified. Parts of this page were drafted with the help of an AI writing assistant. A Europe by Giddo editor checked the facts, corrected the text and approved it before publication. Europe by Giddo holds editorial responsibility for its content.
Where an illustration was generated with AI, the picture carries its own short notice — “Illustration generated with AI. It is not a photograph of this place.” — and links to this page. A link to this page also appears in the footer of every page on the site.
3. Where AI is used
- Drafting. Producing a first draft of a description from research notes and verified sources, which an editor then rewrites, cuts and corrects.
- Editing. Grammar, clarity, consistency of tone, shortening and structuring.
- Research support. Suggesting angles, summarising material, and flagging things worth verifying — never as the source of a published fact.
- Translation support. The site is published in English; assistance may be used when working from source material in another language, and the result is checked by a person.
- Software. Parts of the application code have been written with AI assistance, reviewed by a developer, and tested before deployment. This is disclosed for completeness; it says nothing about the accuracy of any editorial entry.
4. Where AI is not used
- Never to publish. There is no automatic publication path. Publication is a deliberate human action taken by an authorised account, and it is recorded in an audit log.
- Never to invent facts. Opening hours, prices, admission conditions, coordinates, distances, contact details, awards, credits and sources are not generated. They are taken from a verified source or they are omitted. An entry that lacks a fact will say nothing rather than guess.
- Never to fabricate experience. We do not publish invented visit reports, invented recommendations or first-person accounts of visits that did not happen.
- Never to fabricate reviews. There are no user reviews on this site and no generated ones.
- Never to generate photographs of real places. Images of real destinations are photographs, used with their rights and credits recorded. We do not present a synthetic image as a photograph of a real place. Where an illustration is generated, it is recorded as such in our media catalogue with its model, provider and date, and is labelled where it is displayed.
- Never to recreate Giddo. The Giddo character is not generated, redrawn or reinterpreted. Only real, approved Giddo artwork is used.
- Never to make decisions about people. No profiling, no scoring, no automated decisions about visitors — see Your GDPR Rights, Article 22.
5. Who is responsible
LAVALI ONLINE SHOP SRL holds editorial responsibility for all content published on this site, whether or not an AI assistant was involved in producing it. “The AI wrote it” is not an excuse we will ever offer you. If something here is wrong, it is our error, and we want to know: Report Content.
6. How this relates to the law
Article 50(4), second subparagraph, of Regulation (EU) 2024/1689 (the AI Act) requires deployers of an AI system that generates or manipulates text published to inform the public on matters of public interest to disclose that the text has been artificially generated or manipulated. That provision applies from 2 August 2026.
The same subparagraph states that the obligation does not apply where the AI-generated content has undergone a process of human review or editorial control and a natural or legal person holds editorial responsibility for its publication. Our workflow is exactly that: mandatory human review, explicit human publication, and editorial responsibility held by LAVALI ONLINE SHOP SRL.
We therefore disclose voluntarily. We are not claiming a legal duty we do not have, and we are not using the exemption as a reason to say nothing. The exemption explains why no warning label is legally mandated on our editorial pages; our own editorial standard is why the notice in section 2 appears anyway.
Where an AI system generates or manipulates image, audio or video content constituting a deep fake, Article 50(4), first subparagraph, does impose disclosure. We publish no deep fakes. Should a generated illustration ever be published, it will be labelled as generated at the point of display, not only in this policy.
7. How we keep this honest
- Every editorial entity has a full revision history: what changed, who changed it and when, with a checksum on each snapshot.
- Publication, unpublication and restoration are recorded in an audit log tied to the acting account.
- Media assets carry their rights, source, licence and — where relevant — an AI disclosure record naming the model, the provider and the date of involvement. An asset cannot be attached to a published page until it has been reviewed and approved.
- The publication rules are enforced by the software, not by memory: content is draft until a person with the publishing permission changes its state.
8. Tell us when we get it wrong
Verification is a process, not a claim. If you find an inaccuracy, an out-of-date detail or something that reads as though nobody checked it, please use Report Content. Corrections are made in the same place as the error, and the revision history keeps the record.
9. Changes to this notice
This notice is versioned and dated at the top of the page. It will be revised if our practice changes — for example if generated illustrations are introduced, or if the applicable transparency rules change. We will not quietly widen what AI is allowed to do here.