Correctie van beleidsrecords en bewijs
Add ANU, Edinburgh, and CUHK policy records with verified claims, exact presentation patches, and restored CUHK analysis dimensions.
De samenvatting wordt in het Engels weergegeven.
Historische correctie: aantallen en claims horen bij deze versie, niet bij het huidige record.
· v6-target-r7-candidate-20261021
- Université PSL · 7 reviewed records in this revision · Beleidspagina bekijken
- Toegevoegde records: 0; Gecorrigeerde records: 0; Aangehouden / uitgesloten van huidige conclusies: 0
Record revision, scope and original evidence
Original evidence and audience, course and unit scopes are preserved. Tracker correction time is not a policy effective date.
The records below reflect this historical version, not the current policy record or current review conclusions.
2e03224c6de672fbd610712c5d21072279dbf69f1b57cbbb3442d4a56c6704c5
{
"addedClaimIds": [],
"updatedClaimIds": [],
"heldClaimIds": []
}Previous baseline and review records- psl-claim-central-ai-disclosure — Université PSL's Règlement de la scolarité for 2026-2027 (adopted by the Board on 25 June 2026, replacing the June 2025 version with the same AI wording) states that, where an instructor authorises AI-based tools such as ChatGPT, their use must be explicitly disclosed like any source citation, and failure to disclose AI use is treated as plagiarism and sanctioned as such; a disciplinary procedure may follow, with sanctions up to exclusion (not automatic exclusion). The 2026 version applies to programmes operated directly by PSL (such as the Collège Sciences, Humanités et Société and the Paris School of AI); member institutions apply their own regulations, so it does not set course permissions across ENS, Dauphine, Mines Paris and other members.
- psl-claim-research-transparency — Research-integrity guidance published by the Université Paris Dauphine-PSL library recommends that researchers who use generative AI state the tool, its version and the context of use, and notes that prompts and results may also be documented. This is library guidance for researchers at a PSL member institution, not a PSL-wide mandatory rule.
- psl-claim-research-responsibility — The Dauphine-PSL library's research-integrity guidance advises keeping a critical approach to AI tools, not treating them as co-authors, and not using them to create scientific materials that could be falsified. It is advisory library guidance at a member institution, not a PSL-wide administrative prohibition.
- psl-claim-research-data-privacy — The Dauphine-PSL library's research-integrity guidance states that personal data are not meant to be shared with third parties and that generated text should be checked so it is neither plagiarism nor contains personal data. This is member-institution library guidance.
- psl-claim-research-limitations — The Dauphine-PSL library's guidance notes that generative AI undermines research reproducibility, a basic principle of scientific integrity, because models' code and algorithms may be unknown and change regularly. This is a risk statement, not a prohibition on AI use.
- draft-claim-psl-doctoral-charter-genai — PSL's Charte du doctorat (June 2026, section 2.7) sets general recommendations on generative AI for doctoral candidates, some of which are mandatory. Generative AI is not prohibited in doctoral research, but its uses must be transparent, declared and critically assessed, and the candidate remains fully responsible for AI-obtained content. Candidates must insert a dated, signed 'Déclaration doctorale d'intégrité scientifique' (one to two pages) at the start of the thesis, stating which generative AI tools were used, for which tasks, their benefits, limits and risks, and any mitigation. Member institutions may set stricter rules, which prevail.
- draft-claim-psl-ciens-ai-charter — The AI charter of CIENS (an interdisciplinary centre of ENS-PSL) applies to CIENS students, teachers and researchers only. It allows generative AI as occasional help for translation, document exploration, exploratory document analysis or organising work, but prohibits producing false references, developing a research problem and scientific reasoning, writing the text content, and giving AI sensitive, protected or confidential information. Any AI use must be stated in a note or the body of the text, naming the tool and the nature of use; undeclared use may be treated as academic fraud and lead to disciplinary sanctions. These are unit-level rules, not PSL-wide rules.