Ethics Assessment of Prêt-à-LLOD
Are we doing anything wrong in Prêt-à-LLOD? What is the impact of the Prêt-à-LLOD algorithmic decision-making and machine learning in the context of the project pilots?
The Prêt-à-LLOD project partners have been mandated to answer these questions, or at least to further research on the consequences of the execution of this project. But in this inquiry, the Prêt-à-LLOD team is not only acting due to a contractual obligation, but it is also motivated by the awareness of the potentially high impact of this project. The fast development of language technologies, and the growing extension of their application are unleashing fundamental changes in our society, and even if Prêt-à-LLOD represents a tiny contribution to a revolution, the Prêt-à-LLOD team wants to evaluate its impact.
For this regard, 25 people working for the Prêt-à-LLOD people conducted a participatory activity where these questions were systematically analysed. The activity was facilitated by Víctor Rodríguez-Doncel and the UPM team in Vienna in November 2019. Having formed four working groups a canvas-based analysis was made for each of the pilots. The people were arranged in groups of approximately equal size, and were moderated by one person who scribed the discussions. The ADAPT Ethics Canvas (template available online) was used as the guiding thread, and the notes taken by the facilitator were shared in a later stage. A summary and conclusions have been presented in Deliverable D7.2.