Faculty of Science Directive on the Use of

Generative Artificial Intelligence

Version December 2025

 

 

  1. Background

    Generative artificial intelligence (GAI) models have seen spectacular advances. They are now capable of solving complex problems, performing multi-step reasoning, and producing texts of impressive structural and editorial quality. Their rapid improvement is permanently transforming academic practice.

    Although there exist legitimate uses for GAI in an educational setting, it never should be used to compensate for a student's lacking level of understanding. Submitted work must reflect the actual skills of its authors. Furthermore, AI can produce erroneous or fictitious content (hallucinations). Its use also carries risks of unintentional plagiarism and copyright infringement, particularly in the absence of critical validation, and the risk of disclosure of private or sensitive data to third parties.

     

  2. Objectives of the Directive

    The purpose of this directive is to establish basic principles for the use of AI in academic work.

    It seeks to:

    • Encourage the useful, reflective and critical use of AI tools (text, image, code, etc.);

    • Clearly explain the permitted and prohibited uses, based on the choices made in each course;

    • Preserve the transparency, traceability, and integrity of academic work.

    It does not impose strict rules, and instead provides a framework to help everyone make good use of these tools.

     

  3. Scope of Application

    Definitions: For the purposes of this guideline, AI tools comprise any digital tools that, based on instructions (prompts) or text provided, produce new content or rewrite existing content (e.g., text, code, images, audio, summaries, translations).

    Examples: writing assistants (e.g., ChatGPT, Gemini), translation and rewriting (e.g., DeepL Translate/Write), code assistants (e.g., Copilot), image/video/voice generators, synthetic data creation.

    Minor software aids (excluding GAI). Local and deterministic functions that do not add intellectual substance (basic spell checking, automatic closing of parentheses, indentation/formatting, trivial suggestions) do not fall within the scope of GAI.

    Principle of transparency. Whenever a tool produces, rewrites, translates, or summarizes part of the work submitted by a student, its use must comply with the rules of authorization and disclosure set out in this directive.

     

    It applies to all members of the faculty community: students, teachers, in the context of teaching, learning, academic production, and evaluation activities, and concerns the use of tools such as ChatGPT, Gemini, PerplexityAI, DeepSeek, Grok, Claude, DALL·E, Copilot, Midjourney, DeepL (Translate/Write), etc.

     

    Research activities (including scientific work conducted outside of a teaching or assessment context) are not covered by this directive.

    This directive applies to all educational activities organized within the Faculty of Science, whether related to teaching, tutorials, exams, projects, dissertations, or theses. It applies to anyone taking part in these activities, depending on the role in which they are involved (for example, a doctoral student may be concerned both as a student and as a teacher).

     

    It concerns two groups:

     

    Students:

     

    • Persons enrolled in a bachelor's or master's program;

    • Doctoral students, when they are taking courses or enrolled in a doctoral program, excluding their independent research activities. Writing a thesis manuscript and preparing for a thesis defense are considered part of their student activities.

    • Participants in continuing education or micro-certification modules, provided they are participating in a teaching or academic validation activity.

    Teaching staff:

     

    • Anyone who actively contributes to a teaching activity in the Faculty's programs;

    • Supervisors of projects, theses, or laboratory sessions;

    • People who evaluate, supervise, or oversee student work;

    • Thesis advisors.

     

    Note: Teaching status is determined by the educational role assumed, regardless of administrative status.

     

  4. General Principles

    1. Personal work is primarily the result of the intellectual and personal contribution of its authors.

    2. GAI can serve as a tool, but it cannot replace the personal work, analysis, originality, critical thinking, or critical reasoning expected of a student.

    3. All use must be transparent, verifiable, and cited according to the citation rules defined by the instructor or in use in the scientific discipline of the work.

    4. Students must check with their teachers to what extent they are authorized to use AI in order to avoid the dissemination of sensitive data.

    5. Teachers are free to use or not use GAI in their teaching. They may also authorize or prohibit the use of GAI in assessments, provided that the final validation of the results remains the responsibility of the teacher. The permitted and prohibited uses of GAI must be explicitly mentioned in the teaching materials (handouts, Moodle, lesson plans, etc.).

    6. Teachers are also free to define the specific terms and conditions for students to declare their use of GAI. This declaration may be required either generally, for example via a form integrated into Moodle or any other system at the beginning of the course, or on a case-by-case basis, for each assignment, lab session, or assessed activity.

    7. In all cases, a formal declaration of GAI use is required for all bachelor's, master's, or doctoral work, as well as for any academic assessment or production leading to the award of a degree or official title. This declaration must follow the standard template provided by the Faculty.

    8. Sensitive data, such as personal data, must not be disclosed to a public artificial intelligence tool that does not guarantee confidentiality. Similarly, it is recommended that all other confidential data (e.g., unpublished research results or data from theses in preparation) be protected from exposure to a generative artificial intelligence tool. This principle applies regardless of the instructions given by teachers and remains valid in all contexts of use. In addition, teachers who use sensitive data in their courses or supervise work containing such data are required to explicitly remind students of the principles of confidentiality and data protection, in accordance with the spirit of these recommendations.

     

  5. Evolution

    In light of the spectacular advances in GAI tools and their rapid progress, this directive will be revised regularly. Students can submit their feedback via student representatives.