On Friday, November 17th, Etienne, Emma & Rubing organized a work session at l’Ecole Supérieure de Journalisme de Lille. The workshop was part of a three-day program on AI and journalism aimed at providing students with an overview of the technology and its social impacts.

In order to go beyond the usual, prophetic discourses about AI, we introduced the students to the logic of supervised machine learning, Transformers architectures, as well as BERT and GPT models. We then asked them to reflect on journalistic writing by designing indicators and subsequently annotate text.

During the workshop, we presented various tools to massively explore corpora, premised on our Augmented Social Scientist paper and our soon-to-be-released annotation platform Active Tigger – on which they manage to create a supervised classifier that captures the use of « off the record » in French journalism.