Do not use any generative AI tools for administrative or technical tasks without understanding their implications and without knowing exactly what you want to achieve.

How can you do this?Before using generative AI, you must be very clear about which task needs to be performed (drafting minutes, classifying documents, generating a database, etc.) you want to do and how AI can help you with your supervision.

AI makes mistakes or invents data, and may also provide outdated or incorrect regulatory and legal frameworks, which is particularly serious in the administrative sphere.

What can you do?Always treat data or text generated by AI as a first draft, and make it standard practice to verify all information (particularly numbers, legal terminology, protocol names, references to regulations, etc.)with official sources, whether from the university or against the original documents. Never include unverified information from generative AI in a file.

This point is crucial for PTGAS, which routinely handles personal, academic, payroll or recruitment data.

What can you do?Never enter personal, identifying, or sensitive data into any generative AI tool, if its servers are not located within the European Economic Area. Before using a tool, make sure that it complies with the GDPR and that the institution permits its use. If in doubt, fully anonymise the information or use institutional tools that guarantee privacy.

Generative AI can be a management support tool provided it does not replace professional judgement, knowledge of administrative procedures, or responsibility for the case file.

How can you do this?Any use of generative AI involves initial, intermediate and final supervision (from summarising reports to improving the writing of communications). This also means that decisions and support for students or staff must continue to be provided in a personalised way, and must not dehumanise the support by using generative AI responses.