AI renders many university assessments (notably the dissertation) worthless. Let’s return to oral assessment

True progress occurs when technology and the market make available to the many what was once the preserve of the very few: telecommunications, air travel, citrus fruits in winter, works of art, anaesthesia in the delivery room, diabetes, obesity, debt. Whether good or bad, certain things are no longer as rare as they once were. Among these is automatic writing, an invention of the Surrealists who would jot down words in a rush, without reflection, as if possessed, and which – given the rather disappointing results – nobody, or almost nobody, would dream of publishing. Degree theses over the last twenty years have popularised that model: written in a trance by most students, the examiners, rather than reading them, scrutinise them from top to bottom. In practice: running their eyes from top to bottom and focusing only on the central part of the page, because, given the fragments, examining the whole thing doesn’t make much difference anyway. Never in history have there been so many words without an author or readers.

Now that AI exists, the floodgates have opened. Since it is impossible to ban it, every effort is being made to regulate it. The University of Modena and Reggio has made 30,000 licences available to lecturers and students – almost one per person – specifying that ‘AI may be used as a support for understanding complex concepts, developing ideas, carrying out self-assessment activities and reworking information’. But understanding concepts and reworking information is precisely the point of studying. If we delegate this to AI, it is tantamount to encouraging people to stop thinking. Moreover, there is no detection software efficient enough, partly because the distinction between human and artificial writing is now too subtle. So, what should we do? Return to the oral tradition of the Italian university. Asking students to churn out reports is a practice that originated in nineteenth-century German seminars and was subsequently exported to the Anglo-Saxon world, but which, given the times, is proving to be a double-edged sword. Oral exams take longer and are more demanding for everyone, granted. But that is precisely why they are more useful and revealing. As for dissertations, it would be a very good idea to do away with them. Why expect such a complex writing task from young people who can barely read a novel?

Let’s seize the opportunity: instead of asking the two customary questions based on a pile of text written by who knows whom – questions that wouldn’t even stump Pinocchio and would just make the audience yawn – let’s turn the degree examination into a serious and in-depth discussion of a topic agreed upon between the professors and the student. Only in this way will the student’s knowledge of the subject, their ability to argue a case and to respond to objections be put to the test. AI has not yet learnt to break a sweat in front of an examination board. And without butterflies in the stomach and performance anxiety, what sort of university would it be?