Generative AI Teaching Assistants Reshaping Teacher Knowledge Workflows

This study explores the impact of generative artificial intelligence (AI) teaching assistants on teachers' knowledge workflows through a multimodal interaction log analysis framework, combining petri net modeling and causal inference methods. The research reveals that generative AI significantly enhances task efficiency by reducing node compression rates (up to 28.9%) and optimizing process links (up to 35.1%), particularly in structured tasks like resource pushing. It also alleviates teachers' cognitive load, which is evidenced by a drop in subjective scores from 5.9 to 4.8 during interactive feedback stages. Key findings highlight a positive correlation between interaction frequency and structural optimization, with higher AI usage intensifying task compression effects. The study provides actionable insights for integrating generative AI into educational systems while addressing limitations in complex task modeling and behavioral data granularity.