Corpora and AI in Linguistic Tasks: Casing Idiomaticity

This is a presentation given by Petya Osenova (Sofia University) at the CLARIN in University Curricula Workshop, held in Utrecht on 21–22 April 2026. The aim of this talk is to show how information from large corpora, as well as instruments from corpus management systems, can be combined successfully with AI prompts to handle the meaning and distribution of idiomatic expressions. The CLASSLA-web.bg 2.0 (Bulgarian Web) from the CLARIN CLASSLA K-centre has been used for the corpus data, and ChatGPT for the prompting engineering.