Informação sobre inteligência artificial

O que fazem as funcionalidades de inteligência artificial da plataforma, que fornecedores tratam os dados, o que é enviado, durante quanto tempo é guardado e como a sua escola as pode desativar.

Última atualização:

Funcionalidades de IA

Cada funcionalidade, quem a pode usar, os dados que envia e o fornecedor que os trata.

Pseudonymisation

Before every request, the names of known people (students, guardians, teachers and others) are replaced with placeholders such as [ALUNO_1], as are emails, phone numbers, tax and social security numbers, citizen card numbers, IBANs and dates of birth. The real names are put back only in what the user sees; the mapping never leaves the platform. No user identifier is sent to the provider.

Data retention

The questions and who asked them are kept for cost and abuse control for the period set for each school (90 days by default, between 7 and 3650). After that, only the usage figures and the cost remain. A person can ask the school to erase their AI data at any time. Anthropic, as a sub-processor, handles the data under its data processing agreement.

Human oversight and limits

Every answer is marked as generated by AI and should be checked against the sources given. The AI never grades, ranks, admits, proctors or profiles students, and makes no decisions about people: a user always decides what to do with the answer.

How to turn it off

The school can turn off every AI feature, or only the reading of class data and the pictures, in the e-Schooling AI settings, and remove access from user groups in the permissions (AI feature group). Users can also simply not use the AI search: the Google search stays available.

Dúvidas sobre as funcionalidades de IA ou sobre os seus dados? contacte-nos