The Vice Rectorate for Digital Transformation presents the guide for the use of artificial intelligence services at the UPV, with a practical approach that guides the whole university community in the responsible, safe and efficient use of these tools. The guide is adapted to the needs of the UPV and puts special focus on the protection of personal data contained in the documents.
Artificial intelligences of preferential use at the UPV
The following are considered to be of preferred use because they are integrated in the technological ecosystem of the UPV. This means that do not use the data sent or the conversations to train the models and that the information, including personal data contained in the documents, remains protected. within the university’s infrastructure.
| PoliGPT | Copilot | |
|---|---|---|
| What is it? | Proprietary artificial intelligence managed within the UPV. | Artificial intelligence of Microsoft complements PoliGPT with the advantages of the Microsoft license subscribed by the UPV. |
| Who can use it? | It is in testing, only UPV Staff. | All the university community with the educational account. |
| What can you do? | Basic functions such as: generation of text-based answers, searching for UPV and generic information, solving questions and answers, as well as text-based programming and translation tasks, etc. | All the functions provided by PoliGPT and in addition, the generation, reading and extraction of data from documents or images, etc. |
Other artificial intelligences
The rest of artificial intelligences external to the UPV present in the market, such as Gemini, ChatGPT or Claude can only be used with data that is already published, that is, with public information.
⚠️ Send internal or personal data to artificial intelligences external to the university (Gemini, ChatGPT, Claudeetc.) implies that these data are no longer protected.Everything that is shared -conversations, files or information can be used to train these artificial intelligence models.
In the guide you will find when to use each artificial intelligence depending on whether it contains internal data, i.e. unpublished information, and the functionalities of each tool.