Historique du registre
Mises à jour de nos données vérifiées ; elles ne confirment pas un changement de politique de l’université.
Révisions correspondantes : 4
Mise à jour de page
Révision: v6-reading-rankings-ten-20261027Textes ancien et nouveau non comparables.
Shared V6 reading layout and evidence presentation updated. University policy Claims, original evidence, held decisions and snapshot approvals remain unchanged; the institution audit remains PARTIAL.
Éléments concernés : 15 (11 corrigés / ajustés · 4 ajoutés)
Révision: v6-audit-fix-20261014Éléments concernés : 15 (11 corrigés / ajustés · 4 ajoutés)
Confidentialité · Enseignement · 4 autres thèmes
ETH Zurich requires transparency about GenAI use: students must declare which tools they used and for which parts of their work; lecturers must communicate when GenAI use is permitted and make their own GenAI use visible.
ETH Zurich's teaching guidelines instruct students to be transparent about their use of generative AI and to identify which tools contributed to which parts of their work, in accordance with applicable lecturer instructions and citation conventions. For scientific writing and theses, students should agree in advance with their supervisors on authorised AI use and disclosure requirements. ETH Library guidance recognises an exception where a supervisor has expressly agreed that particular AI use need not be separately declared. In the absence of such an agreement, undeclared use may constitute a breach of academic integrity.
Textes ancien et nouveau non comparables.
Violations of GenAI guidelines such as use of unauthorised aids or non-disclosure of their use are subject to disciplinary action under existing performance assessment rules and the declaration of originality.
Violations of GenAI guidelines such as use of unauthorised aids or non-disclosure of their use are subject to disciplinary action under existing performance assessment rules and the declaration of originality.
Textes ancien et nouveau non comparables.
ETH Zurich states that technical recognition of AI-generated output is currently unreliable and will probably remain so; trust in such methods is not appropriate.
ETH Zurich states that technical recognition of AI-generated output is currently unreliable and will probably remain so; trust in such methods is not appropriate.
Textes ancien et nouveau non comparables.
Students are responsible for the content of work they submit. Performance assessments must be conducted independently and personally; GenAI may serve a supplementary role but not replace student efforts.
Students are responsible for the content of work they submit. Performance assessments must be conducted independently and personally; GenAI may serve a supplementary role but not replace student efforts.
ETH Zurich's declaration of originality provides two options: confirming that no generative AI was used, or confirming that authorised generative AI was used and its disclosure agreed with the supervisor. Bachelor's and Master's theses require a signed declaration of originality, while requirements for other written assessments depend on the applicable academic instructions. Doctoral candidates declare independent authorship through the doctoral examination registration process. The permitted use, disclosure format and any exceptions must follow the relevant supervisor and unit requirements.
ETH Library provides Turnitin Originality to eligible professors and lecturers for text-similarity checking, subject to data classification and responsible selective-use guidance. Only public or internal documents may be uploaded. Similarity screening is distinct from AI-authorship detection; ETH Zurich warns that technical detection of AI-generated output is unreliable and should not be treated as conclusive evidence.
Textes ancien et nouveau non comparables.
Lecturers determine whether and how GenAI may be used in their courses and for respective assessments. Teaching materials created with GenAI must be subjected to quality control by the lecturer.
Lecturers determine whether and how GenAI may be used in their courses and for respective assessments. Teaching materials created with GenAI must be subjected to quality control by the lecturer.
Textes ancien et nouveau non comparables.
Students must refrain from disclosing copyrighted, private, or confidential information to commercial GenAI clients unless expressly permitted, and must respect privacy and copyright of content they work with.
Students must refrain from disclosing copyrighted, private, or confidential information to commercial GenAI clients unless expressly permitted, and must respect privacy and copyright of content they work with.
ETH Zurich prohibits the processing of strictly confidential information in external cloud AI services. Confidential information may be processed only in services and configurations explicitly approved for that classification. ETH IT guidance distinguishes Microsoft 365 Copilot Work mode without web search, which is approved for confidential data, from Copilot Chat and web mode, which are approved for internal data. Unassessed AI services must not be used for confidential ETH data or protected personal data.
ETH Library guidance requires researchers to remain responsible for their scientific work and to verify AI-generated information, references and quotations. Generative AI tools cannot be listed as scientific authors or relied upon as authoritative scientific sources. Researchers should agree on permitted AI use with collaborators and follow applicable departmental, funder, publisher and journal requirements. Disclosure of an AI tool's contribution is distinct from citing that tool as evidence for a scientific claim.
Textes ancien et nouveau non comparables.
ETH Zurich recommends Microsoft Copilot, Google Gemini, and NotebookLM for teaching purposes, as they offer data-protected access via ETH accounts where personal data is not used for training models.
ETH Zurich recommends Microsoft Copilot, Google Gemini, and NotebookLM for teaching purposes, as they offer data-protected access via ETH accounts where personal data is not used for training models.
Google Gemini is listed for ETH Zurich in an official university AI tools source. Derived availability: allowed. Derived endorsement type: institutionally licensed or procured.
Google Gemini is officially licensed and provided to ETH Zurich employees and students free of charge through the ETH Google Workspace Cloud Subscription with enterprise data protection (no model training). CISO cloud assessment remains OPEN; Google tools are not approved for confidential or protected personal data. An optional fee-based Google AI Pro license is available for employees.
Microsoft Copilot is listed for ETH Zurich in an official university AI tools source. Derived availability: allowed. Derived endorsement type: institutionally licensed or procured.
Microsoft Copilot is officially licensed for ETH Zurich. M365 Copilot Chat is available free to all members and approved for internal data. M365 Copilot Work mode without web search is approved for confidential data for licensed employees. Uploading strictly confidential data is prohibited.
Google NotebookLM is listed for ETH Zurich in an official university AI tools source. Derived availability: allowed. Derived endorsement type: institutionally licensed or procured.
Google NotebookLM is officially available free of charge to ETH Zurich members through the Google Workspace Cloud Subscription with enterprise data protection. CISO cloud assessment remains OPEN; tools lacking assessment have not been released for confidential or personal ETH data.
Textes ancien et nouveau non comparables.
ETH Zurich advocates a proactive approach to the use of generative AI in educational contexts, emphasising responsible use among students and lecturers.
ETH Zurich advocates a proactive approach to the use of generative AI in educational contexts, emphasising responsible use among students and lecturers.
Éléments concernés : 6
Révision: public-release-20260715-005Éléments concernés : 6
Règles d’utilisation des outils d’IA
Textes ancien et nouveau non comparables.
Google Gemini is listed for ETH Zurich in an official university AI tools source. Derived availability: allowed. Derived endorsement type: institutionally licensed or procured.
Textes ancien et nouveau non comparables.
Google Gemini is listed for ETH Zurich in an official university AI tools source. Derived availability: allowed. Derived endorsement type: institutionally licensed or procured.
Textes ancien et nouveau non comparables.
Microsoft Copilot is listed for ETH Zurich in an official university AI tools source. Derived availability: allowed. Derived endorsement type: institutionally licensed or procured.
Textes ancien et nouveau non comparables.
Microsoft Copilot is listed for ETH Zurich in an official university AI tools source. Derived availability: allowed. Derived endorsement type: institutionally licensed or procured.
Textes ancien et nouveau non comparables.
NotebookLM is listed for ETH Zurich in an official university AI tools source. Derived availability: allowed. Derived endorsement type: institutionally licensed or procured.
Textes ancien et nouveau non comparables.
Google NotebookLM is listed for ETH Zurich in an official university AI tools source. Derived availability: allowed. Derived endorsement type: institutionally licensed or procured.
Preuves ou classification historiques incomplètes ; consultez l’historique.
Ouvrir l’historique de la révisionÉléments concernés : 3 (3 ajoutés)
Révision: public-release-20260614-001Éléments concernés : 3 (3 ajoutés)
Règles d’utilisation des outils d’IA
Google Gemini is listed for ETH Zurich in an official university AI tools source. Derived availability: allowed. Derived endorsement type: institutionally licensed or procured.
Microsoft Copilot is listed for ETH Zurich in an official university AI tools source. Derived availability: allowed. Derived endorsement type: institutionally licensed or procured.
NotebookLM is listed for ETH Zurich in an official university AI tools source. Derived availability: allowed. Derived endorsement type: institutionally licensed or procured.
