Sign in Subscribe
Machine learning

Leveraging Generative AI for Course Learning Outcome Categorization Using Bloom's Taxonomy

Strategic Insights

— 1 min read

Research conducted at King Abdulaziz University has demonstrated the potential of using generative AI, specifically GPT-4, in categorizing course learning outcomes according to their respective cognitive levels within the revised Bloom's taxonomy. The study employed various prompt engineering strategies, including zero-shot, few-shot, chain-of-thought, rhetorical situation, and multiple binary questions, to assess the effectiveness of GenAI. The results showed that the prompt incorporating rhetorical context and domain-specific knowledge achieved the highest classification performance, while the multiple binary question approach underperformed compared to the zero-shot method.

Key Takeaways:

  • The research utilized a dataset of 1000 annotated learning outcomes to evaluate the effectiveness of generative AI in classifying course learning outcomes according to their respective cognitive levels within the revised Bloom's taxonomy.
  • The study found that the prompt incorporating rhetorical context and domain-specific knowledge achieved the highest classification performance, with an accuracy of 82.5% and an F1-score of 85.2%.
  • The research demonstrated moderate to substantial agreement with expert annotations, indicating the potential of leveraging large language models to advance both theoretical understanding and practical application within the field of education and natural language processing.
  • Aditya Johri and Omaima Almatrafi contributed to the research, with Almatrafi serving as the lead author.
  • The study highlighted the potential of using generative AI to streamline educational processes, improve student outcomes, and enhance teacher training.

Statistics:

  • 1000 annotated learning outcomes were used in the dataset.
  • The accuracy of the prompt incorporating rhetorical context and domain-specific knowledge achieved 82.5%.
  • The F1-score of the same prompt engineering strategy reached 85.2%.
  • The study utilized GPT-4 as the generative AI model.
  • The Cohen's kappa value for the prompt-based approach was 0.81, indicating substantial agreement.

Sources:

  • Leveraging generative AI for course learning outcome categorization using Bloom's taxonomy. Computers and Education: Artificial Intelligence, 2025, 8():100404. Publisher: Elsevier.
  • https://doi-org.sdpl.idm.oclc.org/10.1016/j.caeai.2025.100404

Read more