Using Chat GPT to Clean Qualitative Interview Transcriptions: A Usability and Feasibility Analysis
ZW Taylor
Abstract
<i>One of the major inefficiencies in qualitative research is the accuracy and timeliness of transcribing audio files into analyzable text. However, researchers may now have the ability to leverage artificial intelligence to increase research efficiency through Chat GPT. As a result, this study performs feasibility and accuracy testing of Chat GPT versus human transcription to compare accuracy and timeliness. Results suggest that by using specific commands, Chat GPT can clean interview transcriptions in seconds with a &lt;1% word error rate and near 0% syntactic error rate. Implications for research and ethics are addressed.</i>
Topics & Concepts
UsabilityComputer scienceQualitative analysisWorld Wide WebQualitative researchHuman–computer interactionSociologySocial scienceArtificial Intelligence in Healthcare and EducationTopic ModelingInterpreting and Communication in Healthcare