The document, aimed at students, faculty, researchers, and technical-administrative staff, acknowledges the potential of these technologies to support learning, research, and service efficiency, while emphasizing the importance of transparency, rigor, and responsibility. Key principles include protecting privacy, preventing bias, and preserving critical thinking, with a warning about GAI systems potentially producing incorrect information and the need to verify results.
In teaching, each course unit may define the conditions for GAI use, from free use to total prohibition, based on pedagogical objectives. The importance of diversifying assessment methods, valuing the learning process and academic work authenticity, is also highlighted. In scientific research, AI systems cannot be considered authors or co-authors of publications; scientific responsibility remains with researchers, who must verify GAI-generated content and transparently declare its use when relevant to the presented results.
The guidelines warn against introducing personal data, confidential information, or unpublished research findings into open platforms, encouraging the use of secure institutional solutions aligned with applicable European legislation, including the GDPR. For administrative services, GAI can assist with tasks like summarizing, translating, and document searching, but decisions affecting individual rights, such as admissions, scholarships, hiring, or professional progression, must remain human-led.
The document also outlines plans for training the academic community, monitoring mechanisms, and sharing best practices among schools and research units. Through these recommendations, ULisboa aims to foster responsible integration of Generative AI in higher education, balancing technological opportunities with human responsibility and academic and scientific integrity.
