Research Summary: A multifactor model for assessing programming knowledge based on the synchronization of AI analytics and expert evaluation

The proliferation of generative AI tools, particularly in programming, poses significant challenges to traditional assessment methodologies in higher education. A study involving 187 undergraduate students across three universities in Kazakhstan investigated a multifactor AI-assisted assessment model designed to fairly evaluate student knowledge in programming courses, particularly when AI support is utilized. The research aimed to determine the alignment of this model with expert instructor assessments.

Why it matters

The integration of AI into educational and professional practices necessitates the re-evaluation of assessment paradigms to maintain validity and fairness. Developing robust assessment models that account for AI assistance is crucial for ensuring the integrity of qualifications and the accurate measurement of human competencies. This impacts the quality of future workforces and the credibility of educational institutions.

Key insights

• Generative AI tools introduce new methodological challenges for assessment validity in higher education, especially concerning objective evaluation of student knowledge when AI can generate code.
• A multifactor AI-assisted assessment model is proposed to address the fair evaluation of AI-supported learning practices within programming courses.
• The study compared traditional one-dimensional assessment approaches, focusing on functional correctness, with evaluations produced by instructors and three large language models (GPT, Claude, Gemini).
• The research involved 187 undergraduate students from three universities in Kazakhstan, indicating a practical application and testing ground for the proposed model.

Source

Frontiers in Education — https://www.frontiersin.org/articles/10.3389/feduc.2026.1996990