The use cases described do not fall within the scope of unacceptable risk or high-risk AI systems pursuant to the European Artificial Intelligence Regulation (AI Act).
GenAI must be used exclusively to support activities, always under human supervision, avoiding the automation of decisions. The principle is particularly relevant when the use of GenAI directly or indirectly affects individuals or groups.
Activities involving the following are therefore always excluded:
- profiling, monitoring or evaluation of people, understood as the attribution of individual judgements, scores, classifications or outcomes;
- decisions concerning persons, understood as determinations, outcomes or acts that produce legal effects or have a significant impact on the rights, freedoms or opportunities of the persons concerned.
When use cases involve the use of personal data and protected data in prompts and generated content, the relevant rules and obligations apply in full. In particular, everyone is required to:
- adhere to the guidelines and operational instructions defined by the University, in order to reduce risks to people;
- implement, when required and also with the support of the Data Protection Officer (DPO), the obligations laid down by the legislation (e.g. legal basis, impact assessments, privacy notices, management of rights and security measures).
List of use cases
For each use case, consistency with the Policy, the main points to consider and operational suggestions for informed and responsible use are indicated.