Making generative AI an ally in academia

In recent years, Generative Artificial Intelligence (AI) has sparked both curiosity and concern, also within academic circles. Giulia D'Agostino, a doctoral assistant at the Faculty of Communication, Culture, and Society at Università della Svizzera italiana (USI), discussed the potential applications of generative AI in education during an interview on "Alphaville," which aired on Rete Due (RSI).

On October 31, 2024, Giulia D'Agostino delivered a seminar titled "How to Teach Responsible Use of Artificial Intelligence Tools for Linguistic and Research Tasks" at the USI-SUPSI Campus. This event was part of a series of initiatives promoted by USI and SUPSI in collaboration with the University of Lucerne, focusing on the use of generative AI in higher education. She explored the topics discussed in her recent radio talk during the seminar.

Generative AI is a relatively new technology, but the concept of artificial intelligence has existed since the 1950s and encompasses much more than just text generation. As Giulia D'Agostino explained on Rete Due, the development of models like ChatGPT has brought AI into the spotlight, making it accessible to the general public. "We all tried to use the tools made available by generative AI in every possible way as it seemed like an intriguing game with numerous possibilities. However, from my perspective, these efforts lacked the reflection and critical thinking that should define every human action. Therefore, we should ask ourselves how these tools function, why they operate in a particular way, and assess the quality of the responses they generate," commented Giulia D'Agostino.

Using generative AI tools responsibly, according to the USI doctoral student, involves maintaining a critical perspective on the answers provided. "Models like ChatGPT can be very useful in creative processes, such as brainstorming or summarising the contents of a text. However, it is important for us to oversee these processes. We need to have a general understanding of the material so we can verify that the models have not fabricated any information. While it is beneficial to explore creative and supportive uses of generative AI, we should not trust it without question."

In university settings, there is often a concern that models like ChatGPT facilitate plagiarism or lead to students no longer learning the content of lectures. "I don't think we should have a coercive approach," said Giulia D'Agostino, "those who want to cheat will always find a way to do so. However, this situation gives us an opportunity to reflect on what is truly important to teach our students, perhaps not mnemonic notions, which can easily be found online. We should emphasise the importance of reasoning about the content and developing a critical mindset."

The path proposed in the seminar is somewhere between uncritical enthusiasm and blocking fear: "Polarisation offers no benefits. One should not be uncritically enthusiastic, mistaking generative AI for true intelligence. Conversely, one should not fear these tools; they can be beneficial or harmful depending on their use. It's important to recognise the potential of these technologies and learn how to utilise them effectively.".

Giulia D'Agostino's interview with "Alphaville" (Rete Due - RSI) is available at the following link (Italian only).