AI transparency
In force as of 10/09/2026.
This page describes how artificial intelligence is used on the E-Ticket platform: which systems are involved, on which content, and what stays a human decision.
It answers the transparency obligations of Regulation (EU) 2024/1689 of 13 June 2024 laying down harmonised rules on artificial intelligence (the "AI Act"), whose Article 50 has applied since 2 August 2026. E-Ticket is a deployer of general-purpose AI systems provided by third parties: the E-Cosplay association develops none, trains none and places none on the market.
1. The AI systems in use
| System | Provider | What it does here |
|---|---|---|
| Claude | Anthropic — https://www.anthropic.com/claude | Assistance in writing the platform's texts and its code |
| Nemotron, via NIM | NVIDIA — https://build.nvidia.com | Draft event descriptions, offered to organisers in their own space |
No other system is called by the platform. No machine translation service is used.
2. How it works, content by content
Event descriptions. An organiser may ask for a suggested description based on the details they have already entered: title, date, venue, town, points to mention. The text lands in their own form field, as a draft: they read it, correct it, complete it, then save it — or discard it. Nothing is published without their action. The model is instructed to invent no facts: no programme, no guest, no price, no schedule that is not in the details provided.
Declaration by the organiser. When creating or editing an event, the organiser declares whether they used AI for the title, the description or the image. That declaration is theirs: the platform does not guess, it asks. When it is ticked, the public event page carries a transparency notice, under the organiser's editorial responsibility.
Platform texts and legal pages. Written with the assistance of Claude, then reviewed and decided upon by the E-Cosplay association. The legal pages were drafted on the basis of documentary work; they do not constitute professional legal advice.
Platform code. Written largely with the assistance of Claude, under human review. This concerns how the site works, not its editorial content.
Translations (French → English). They are human, without exception. Bilingual content comes from reviewed, versioned translation files; transactional emails exist in two hand-written versions, picked according to the recipient's language. Contractual and accounting documents — organiser terms, terms of sale, invoices, tickets — are authoritative in French: a machine-translated contract would carry no weight, and a mistranslated refund or liability clause cannot be walked back.
Photos and images. The platform generates no images. Event visuals are uploaded by organisers, who declare whether they were produced by an AI. No synthetic image is presented as a photograph by the association.
3. What is never handed to an AI
- No automated decision about a person. No scoring, no ranking, no profiling, no access decision taken by a model. Opening an organiser account, a suspension, a refund arbitration: those are human decisions.
- No AI at the gate. A ticket is validated by a cryptographic key and a QR code scan, not by a model. No biometric recognition, no face or voice analysis, neither at the entrance nor anywhere else.
- No emotion recognition, and no content misleadingly depicting an identifiable real person: no deepfake, no cloned voice, no synthetic photo passed off as genuine.
- No personal data of a buyer or an organiser is sent to a model. What is sent to generate a description is an event's public details — title, date, venue — not account data, not orders, not customer addresses.
On that basis, none of the uses described here falls under the high-risk AI systems of Annex III to the Regulation, nor under the prohibited practices of its Article 5.
4. Human review and editorial responsibility
Any text produced by automatic generation goes through human review before publication. For event descriptions, that review is the organiser's, who remains the author of their page and responsible for it. For the platform's own texts, it is the E-Cosplay association's.
That review is not a formality: models invent plausible but false details — a time, a guest, an activity that will not take place. On a ticketing platform, an error of that kind is paid for in disappointed visitors and refund requests. The details that commit anyone — dates, venue, prices, access and refund terms — are entered by hand and never produced by a model.
5. How content is flagged
- A transparency notice appears at the foot of an event page as soon as its organiser has declared AI use for the title, the description or the image.
- A link to this page appears in the footer of every page of the platform.
- Purely functional content — interface labels, forms, transactional emails — carries no notice: it is hand-written, and the permanent footer link is enough.
6. Data sent to providers
Texts sent for generation pass through the providers named in point 1 and are therefore processed on their infrastructure. Their use of that data falls under their own terms and privacy policies, which it is for you to consult through the links in the table above.
The personal data of buyers and organisers — accounts, orders, tickets, messages — is processed as described on the Personal data page, and is not used to feed a model.
7. Questions, corrections, reports
A factual error, content that seems wrongly flagged, a question about how this works: contact@e-cosplay.fr. Every report is examined and, where founded, acted upon.
For an event's content, the organiser is its author: the event page says how to contact them. The association steps in where the report concerns how the platform works, or manifestly unlawful content (see the legal notice).
Last updated: 10/09/2026.