In a global context where the massive digitization of linguistic productions is transforming our relationship to knowledge, language, and discourse, artificial intelligence (AI) is emerging as a key driver of epistemological change. Contemporary linguistics can no longer remain impervious to these upheavals. It is now being redefined through interdisciplinary approaches that combine language sciences, natural language processing (NLP), digital humanities, textometry, and corpus engineering (Silberztein, 2019; Lebart, Salem & Berry, 1998).

This interdisciplinary convergence is profoundly reshaping the ways we think about, produce, and transmit linguistic knowledge. Indeed, the creation of digital corpora – whether oral, written, multimodal, literary, or specialized – is simultaneously a scientific, political, and technical act. As Habert, Nazarenko, and Salem (1997) have emphasized, corpus construction is always the result of methodological and ideological choices that guide the selection and annotation of data. In this regard, Pearson (1998) points out that compiling a corpus is not a neutral act but it involves decisions that affect the representativeness and bias of the data collected. The criteria for selection, annotation, and structuring thus raise a fundamental question of representativeness: which linguistic varieties are valorised? Which voices are ignored? Which cultures are rendered invisible in today’s corpora?

Moreover, the rapid rise of AI in both scientific and media spheres is accompanied by an accelerated evolution in terminology. Concepts such as deep learning, neural networks, transformers, and language models are reshaping traditional categories. In this perspective, Gaudin (2003) asserts that terminology is closely linked to social and scientific dynamics and it reflects the evolution of knowledge and the institutionalization of disciplines.

This transformation is also clearly displayed in language education. Corpora – and especially oral ones – serve as valuable resources for language learning: collocational and phraseological patterns, microsyntax of lexical entries, study of specialized languages, and text typologies (Chambers, MélangesCRAPELno. 31). When enriched and analysed using digital or AI tools, these corpora enhance their pedagogical potential: automatic extraction of linguistic phenomena, discursive contextualization, and adaptability for different learner profiles. However, their integration in the classroom must remain critical: which tools for which audiences? What role should be given to human interpretation? How can we ensure alignment with the diversity of learners?

Finally, the advent of AI in the field of language sciences calls for critical reflection on the knowledge it generates. Beneath the power of statistical models, what do algorithmic biases, automatic labelling processes, or training data reveal about our representations of language? As Ducel, Névéol, and Fort (2022) argue, questions related to fairness and the absence of stereotypical bias are becoming important quality criteria to consider in natural language processing applications. From this perspective, AI-enhanced linguistics requires an epistemological examination of its foundations.

It is within this spirit that the upcoming conference shall be a space for exchange, collaboration, and critical inquiry into linguistic, terminological, and pedagogical practices in the age of artificial intelligence. The conference shall bring together researchers, teachers, doctoral students, and practitioners to engage in cross-disciplinary analyses, share field experiences, and reflect on future orientations.