Israeli Teachers Prefer In-Class Exams Over AI-Aided Homework to Ensure Authentic Learning

A recent study by the Holon Institute of Technology (HIT) and the Hebrew University reveals how artificial intelligence (AI) is reshaping assessment methods in Israeli middle schools. Conducted by Dr. Meital Amzalag and Prof. Gila Kurtz from HIT, alongside Omri Hadar and Prof. Yifat Ben-David Kolikant from the Hebrew University, the research is based on semi-structured interviews with 24 Israeli middle school teachers. Nearly 90% of these educators reported restricting AI use in assignments to better verify students' genuine knowledge.

Teachers expressed a clear preference for in-class tests, oral presentations, and classroom work over homework projects when they feared AI might undermine the learning objectives. This approach, termed "pedagogy of blocking" by the researchers, involves evaluating whether AI use advances learning goals or allows students to bypass essential skills. For example, a literature teacher who incorporated AI in some learning phases considered replacing some homework projects with exams to ensure students demonstrate their own understanding.

The study also highlights teachers’ nuanced attitudes toward AI. Some, like a math teacher with a computer science background, worry that AI use could hinder the development of critical thinking skills. Others use AI extensively for language practice but limit it when independent writing is required. The research identified four guiding pedagogical patterns among teachers: fostering higher-order thinking, maintaining classroom authority, ensuring knowledge accuracy, and adapting teaching to student needs.

Despite these insights, many teachers feel uncertain and burdened by the responsibility of integrating AI effectively. Fifteen out of 24 interviewees voiced dilemmas about adapting teaching to the AI era, 17 felt the decision-making responsibility largely rests on them, and 11 described facing these challenges mostly alone. The researchers emphasize that teacher training must go beyond technical skills to help educators discern when AI genuinely enhances learning and when it substitutes the cognitive effort students should make.

The study, published in Teaching and Teacher Education, is qualitative and not intended to represent all Israeli teachers but to identify prevailing pedagogical thinking patterns regarding AI use in education.