Generative AI-Driven Personalized Nudges

Nudging is a concept derived from behavioral economics that subtly influences individuals’ behaviors toward beneficial outcomes. In education, nudges have effectively guided students toward positive academic performances. However, nudging faces individual differences and contextual appropriateness constraints, which limit their usefulness in intervening behaviors. To address these limitations, this study leverages insights from behavioral economics and taps into the capabilities of generative artificial intelligence (GenAI) to create personalized nudges delivered via mobile-based web applications. These nudges are tailored to each student’s needs and encourage informal learning engagement. A 14-day diary study with 32 university students examined how GenAI analyzed their routine activities and delivered nudges at optimal times. The effectiveness of these nudges was evaluated using latent profile analysis, which identified two student behavior types, with most students responding positively to the learning schedule. Ethical considerations were applied in the nudge design and implementation to minimize disruption to students’ routine academic activities, mitigate GenAI biases, and ensure privacy protection. Overall, the results suggested that GenAI-driven nudges are crucial in providing personalized and timely interventions that positively influence students and support their learning. This demonstrates the innovative use of GenAI for educational purposes and its potential to promote personalized learning.