AI-driven, human-supervised.
I openly use artificial intelligence to design, maintain and enrich my websites. This charter sets out the commitments that guide that use.
1. Identified human responsibility
These websites are designed and led by Rudy Desnos. I take responsibility for their direction, content and use of artificial intelligence. AI may carry out tasks; it does not assume my responsibility or determine the editorial direction on its own.
2. What AI is used for
AI may contribute to monitoring, documentary research, structuring and writing, translation, data analysis, illustrations, development and maintenance. Its role varies according to the content and the website.
3. Sources before claims
An AI response is not evidence. Facts, quotations, figures and dates must be linked to identifiable sources. Established facts, analysis, estimates and hypotheses are distinguished. Unconfirmed information is identified as such or omitted.
4. Supervised automation
Data collection, classification and updates may be automated within a defined, supervised scope. This does not mean that every update is reviewed manually. New editorial content that commits the website requires human approval; an anomaly or uncertainty calls for verification before publication.
5. Clear transparency
A substantial AI contribution is disclosed alongside the relevant content: writing assistance, AI-assisted translation, AI-assisted analysis or an AI-created illustration. This charter describes the shared framework; it does not replace the disclosures needed on individual publications.
6. Images that do not fabricate facts
Drawn portraits and illustrations created or transformed with AI are artistic representations. They are not documentary photographs. No event, interview, testimony, statement or endorsement by a real person may be invented and presented as authentic.
7. Data with understandable origins
Indicators must state their source, date and, where necessary, calculation method and limitations. An estimate is not an observed measurement; missing data does not mean zero. Fictional demonstration data must be identified.
8. Respect for people and contributions
AI use must take privacy, confidentiality and creators’ rights into account. Confidential documents and personal data must not be shared with the tools used without appropriate precautions. In forums, moderation assistance must not impersonate a real participant; disputed decisions must be open to human review.
9. Translations that preserve meaning
An AI-assisted translation is distinguished from the source text. Quotations retain their origin and their meaning must be preserved. When in doubt, readers are invited to consult the original.
10. Correctable errors and an evolving charter
AI can produce errors, misleading connections or bias. Human supervision is also fallible. A reported error must be reviewed and corrected; a significant correction must be made visible. This charter will evolve with our practices while retaining the same human responsibility.
On this website
AI proposals remain subject to business rules and human responsibilities. Actions that commit a person or organisation require authorisation appropriate to their scope.
To report an error, identify the page, the relevant passage and, if possible, a source that supports the correction, using the contact methods shown on the website.
Rudy Desnos
Shared charter — version of 4 October 2026.