This study investigates how Whisper handles interactional phenomena in spontaneous Italian conversation, focusing on backchannels, repairs, and filled pauses. We compare standard Word Error Rate (WER) optimization with a decoding strategy that explicitly rewards the preservation of interactional events. Results show that decoding choices have limited impact on overall accuracy, while recognition remains strongly phenomenon-dependent, suggesting structural limitations in the handling of interactional phenomena, with systematic linearization of repairs and frequent suppression of short conversational items.

This paper analyses the implementation of Automatic Speech Recognition (ASR) into the transcription workflow of the KIParla corpus, a resource of spoken Italian. Through a two-phase experiment, 11 expert and novice transcribers produced both manual and ASR-assisted transcriptions of identical audio segments across three different types of conversation, which were subsequently analyzed through a combination of statistical modeling, word-level alignment and a series of annotation-based metrics. Results show that ASR-assisted workflows can increase transcription speed but do not systemically improve accuracy or prosodic annotation quality. Improvements appear to depend on multiple factors, including workflow configuration, conversation type and annotator experience. These findings are therefore yet not generalizable and highlight the complex interplay between transcription expertise, data type and workflow design. Despite current limitations, ASR-assisted transcription, potentially when supported by task-specific fine-tuning, could be integrated into the KIParla transcription workflow to accelerate corpus creation without compromising linguistic and annotation quality. More broadly, this work underscores the potential of semi-automatic transcription for corpus building, especially in complex settings involving multiple speakers and spontaneous, conversational data.