Athanasios Karasimos, Evangelia-Antonia Efstratiadou, Christos Papatzalas & Ilias Papathanasiou
This paper proposes a systematic framework for integrating Large Language Models (LLMs) into corpus-based teaching for Speech-Language Pathology, Computational Linguistics, and Clinical Neurolinguistics programs. Traditional corpus analysis, though essential, presents significant pedagogical challenges, including being time-intensive and requiring specialized technical expertise, especially with atypical language data like aphasic speech. The emergence of LLMs offers a transformative opportunity to overcome these barriers. The systematic framework introduces practical educational modules, validated annotation workflows, and assessment strategies, exemplified through the Greek CACLA corpus. Crucially, the approach advocates for using LLMs as "cognitive partners" to handle routine tasks, allowing students to focus on higher-order analysis, clinical interpretation, and the critical evaluation of AI outputs. This method aims to democratize sophisticated linguistic analysis while ensuring students develop necessary critical thinking capacities and technological literacy.
Key words: Large Language Models, Computational Linguistics, Speech-Language Pathology, Clinical Neurology, Aphasia, Corpus Methodology