Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Improving Medical Assessment Validity with Generative AI: Lessons from Human-AI Interaction in Bloom's Taxonomy

Rocicleide de Lima Lopes1, Marcos Kubrusly1, Juliana Silva Arruda1

  • 1Christus University, R. João Adolfo Gurgel, 133 - Cocó, Fortaleza, CE 60190-180 Brazil.

Medical Science Educator
|July 13, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Associations between maternal and child anemia and development outcomes: evidence from a large cross-sectional population-based study.

BMC public health·2026
Same author

Prevalence and factors associated with unplanned pregnancy among women of childbearing age in Ceará, Brazil: a cross-sectional study with population representativeness.

Reproductive health·2026
Same author

Beyond the Cognitive: The Role of Social and Psychological Skills in Medical Academic Performance in the Context of Active Learning.

Medical science educator·2026
Same author

Triple burden of malnutrition in children and mothers from Ceará, brazil: a cross sectional study.

Journal of health, population, and nutrition·2025
Same author

Factors associated with interrupted pre-exposure prophylaxis (PrEP) in a Brazilian cohort: a brief communication.

AIDS research and therapy·2025
Same author

Musculoskeletal pain associated with common mental disorders among bus terminal workers in Brazil.

Revista brasileira de medicina do trabalho : publicacao oficial da Associacao Nacional de Medicina do Trabalho-ANAMT·2025

Generative AI tools can help medical educators accurately classify assessment items, improving clinical reasoning skills. While AI has a slight flaw, it outperforms un-aided human judgment, boosting assessment validity.

Area of Science:

  • Medical Education
  • Artificial Intelligence in Healthcare
  • Cognitive Science

Background:

  • Bloom's Taxonomy is crucial for higher-order clinical reasoning but is underutilized by faculty.
  • Current assessment methods often prioritize memorization over clinical reasoning.
  • This gap hinders the development of essential medical competencies.

Purpose of the Study:

  • To develop and validate a Generative AI tool for classifying medical assessment items.
  • To evaluate the impact of this AI tool on faculty's classification accuracy.
  • To enhance the alignment of assessments with higher-order cognitive skills.

Main Methods:

  • A two-phase experimental design was employed.
  • Phase 1: Trained and validated a ChatGPT-4 model against a gold standard of 200 items, tested on 100 items.
Keywords:
Assessment validityBloom's TaxonomyGenerative AIHuman-AI interactionMedical education

Related Experiment Videos

  • Phase 2: Assessed experienced medical faculty's adherence to the AI tool's recommendations on 100 items.

    • Main Results:

    • The AI model achieved 95.0% accuracy and "very good" inter-rater agreement (Kappa=0.85) in Phase 1.
    • Faculty adhered to AI recommendations 75.2% of the time in Phase 2.
    • When faculty overrode the AI, their decisions were correct only 29.4% of the time, indicating overconfidence.

    Conclusions:

    • Generative AI, despite minor limitations, is a valuable partner for medical educators.
    • The AI tool offers more reliable classification than unaided human judgment.
    • It can improve assessment validity, support faculty, and address pedagogical gaps in higher-order thinking skills.