AI usage charter

Regulation (EU) 2024/1689 of 13 June 2024, the AI Act, has applied in principle since 2 August 2026. It does not regulate artificial intelligence in general: it classifies uses by level of risk.

Education and vocational training are among the most closely regulated areas of the Act. Point 3 of its Annex III classes as high-risk those systems used to decide on a person's access or admission, to evaluate their learning outcomes, to determine the level of education they may access, or to monitor their behaviour during a test. The Act goes further for one specific use: Article 5 outright prohibits systems intended to infer a person's emotions in an education institution.

Synodeo uses artificial intelligence. The question is therefore not whether the Act applies - it does - but where the line runs. This charter draws it.

What Synodeo does with artificial intelligence

Synodeo uses a single AI tool: Claude, published by Anthropic. No other system is in use to date. It serves the teaching team and management in the school's internal work: designing, structuring and reviewing course content and frameworks; coordinating internal work - plans, schedules, clean-ups, documentation summaries; drafting and correcting administrative, legal or editorial documents; preparing materials and working templates.

In every case, a person reviews, decides and signs. Teaching and editorial responsibility remains entirely human.

What Synodeo does not do

This is where compliance is decided. No AI system is involved, to any degree, in: the admission of a candidate or their assignment to a program; the positioning assessment, the exemptions granted and the determination of the syllabus to be covered; the evaluation of assignments, marking, grading, the final review and the decision to award or withhold a diploma; the monitoring of people in training, behaviour detection, remote exam proctoring; emotion recognition or biometric categorisation; any chatbot or automated assistant on synodeo.com or app.synodeo.com.

The first four exclusions correspond, point by point, to the four cases in point 3 of Annex III. The fifth is a prohibited practice under Article 5. The sixth removes the information obligation of Article 50, since no direct interaction with a system exists.

Every decision concerning you is taken by a person, on the basis of your file, your interview and your work. There is no automated decision-making at Synodeo within the meaning of Article 22 GDPR: human involvement is the rule, not the remedy.

No personal data is sent to an AI system

This is a written internal rule, not an intention. No name, no contact details, no identifying element relating to a candidate, a person in training, a former learner, a contributor or a third party is entered into an AI system. The same applies to the accompaniment situations set out in assignments, which never leave the platform. Where a working document contains identifying elements, they are removed beforehand.

Three consequences follow, and they are verifiable: Anthropic appears neither as a processor nor as a recipient in Synodeo's record of processing activities, because no personal data is sent to it; no transfer outside the European Union takes place on that account; no data protection impact assessment under Article 35 GDPR is required on that basis.

Synodeo's exact role under the Act

The Act distinguishes the provider, who develops a system or places it on the market under their own name, from the deployer, who uses it under their own authority. Synodeo is a deployer, and only a deployer: it has developed no system, puts its name on none, substantially modifies none and makes none available to third parties. The provider of the model it uses is Anthropic, who bears the Chapter V obligations on general-purpose AI models.

As it uses no system classed as high-risk, Synodeo is not subject to the obligations of Article 26 - designated human oversight, log retention, fundamental rights impact assessment. This situation is the result of a choice, not of chance, and the last section sets out what would change it.

AI literacy

Article 4 of the Act, applicable since 2 February 2025 and enforced since 2 August 2026, requires deployers to ensure a sufficient level of AI literacy among their staff and anyone acting on their behalf. Synodeo meets this through this charter, brought to the attention of everyone working for the school; through a written usage rule prohibiting the entry of personal data, restated before any assignment; and through systematic human review of anything intended for a learner or for the public, which requires and maintains an understanding of the tool's limits - approximations, plausible but false statements, absence of verified sources.

Transparency on published content

The second subparagraph of Article 50(4) requires a deployer publishing AI-generated text to inform the public on matters of public interest to disclose it - except where the content has undergone human review and a person holds editorial responsibility for its publication. Content published by Synodeo falls within that exception: it is reviewed, decided upon and published under the editorial responsibility of the publication director. Synodeo nonetheless chooses to say so, because a school that demands academic integrity from its learners cannot be discreet about its own.

The internal rules require every person in training not to present as their own any work produced by an AI system. This charter is the counterpart of that requirement, on the school's side.

What would trigger a revision of this charter

Each of the following would move Synodeo into a different regime and requires revising this charter beforehand, never afterwards: automated support for the positioning assessment, or assisted marking of assignments, or remote exam proctoring - three high-risk systems under point 3 of Annex III; analysis of attention or engagement during video sessions - a prohibited practice under Article 5; a chatbot on the site or the platform - information obligation under Article 50(1); sending learner data to an AI system - processing under Article 28 GDPR, an update to the record of processing and a likely impact assessment.

Your rights and your questions

Any question about the use of artificial intelligence at Synodeo, and any request for an explanation of a decision concerning you, should be sent to contact@synodeo.com. A reply is provided within fifteen days. You may also refer the matter to the CNIL as regards your personal data, and to the national market surveillance authority designated under the Act as regards artificial intelligence.

Last update

Charter established on 17 August 2026. It is revised whenever the tool, the use or the scope changes, and at least once a year.