Scholarlink

In May 2024, 171 pupils in Years 7 and 9 (ages 14–16) in Latvia voluntarily participated in a structured written life interview about their past experiences, current situation, and plans for their moral development. This paper presents the lessons learnt from the qualitative analysis of these interviews, combining AI and computer-assisted methods. The research question explored in this paper is: How useful and trustworthy was the AI tool for qualitative analysis of interviews? The thematic analysis of the data followed three stages: summarisation, frequency analysis, and retrieval of relevant quotations. Each stage involved a combination of computer-assisted methods (using AQUAD 8e) and AI (ChatGPT-4o). The results from both methods were compared to assess the AI’s potential and trustworthiness in summarising, counting, and retrieving information. The AI tool was most useful and quite trustworthy for summarising pupils’ reported opportunities (e.g., family kindness), challenges (e.g., negative influences, laziness), and strategies (helping others, empathy) guiding their moral development (stage 1). The AI tool had a very low trustworthiness for counting the most frequent codes in the data (stage 2), often missing the textual context in Latvian. The AI tool was useful and moderately trustworthy for selecting significant quotations according to predetermined topics (stage 3), but human-driven verification was still necessary. This study shows that the AI tool ChatGPT-4o can notably accelerate qualitative data analysis, but, depending on the task, human oversight is still necessary for ensuring trustworthiness.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Using Artificial Intelligence for Qualitative Analysis of Adolescents’ Interviews About Moral Growth

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Irina Maslo,
  • Manuel Joaquín Fernández González,
  • Svetlana Surikova,
  • Andrejs Mūrnieks

摘要

In May 2024, 171 pupils in Years 7 and 9 (ages 14–16) in Latvia voluntarily participated in a structured written life interview about their past experiences, current situation, and plans for their moral development. This paper presents the lessons learnt from the qualitative analysis of these interviews, combining AI and computer-assisted methods. The research question explored in this paper is: How useful and trustworthy was the AI tool for qualitative analysis of interviews? The thematic analysis of the data followed three stages: summarisation, frequency analysis, and retrieval of relevant quotations. Each stage involved a combination of computer-assisted methods (using AQUAD 8e) and AI (ChatGPT-4o). The results from both methods were compared to assess the AI’s potential and trustworthiness in summarising, counting, and retrieving information. The AI tool was most useful and quite trustworthy for summarising pupils’ reported opportunities (e.g., family kindness), challenges (e.g., negative influences, laziness), and strategies (helping others, empathy) guiding their moral development (stage 1). The AI tool had a very low trustworthiness for counting the most frequent codes in the data (stage 2), often missing the textual context in Latvian. The AI tool was useful and moderately trustworthy for selecting significant quotations according to predetermined topics (stage 3), but human-driven verification was still necessary. This study shows that the AI tool ChatGPT-4o can notably accelerate qualitative data analysis, but, depending on the task, human oversight is still necessary for ensuring trustworthiness.