A new study published on arXiv analyzes learner interactions with an LLM-powered writing chatbot named Penny. Using Transition Network Analysis on over 4,500 sessions and 21,000 interactions, researchers identified two primary engagement patterns: a 'Revision Loop' for direct error correction and a 'Chat Loop' for extended dialogue. The study found that learner proficiency significantly influences these patterns, with higher-proficiency learners engaging more in negotiation and lower-proficiency learners relying on repetitive feedback cycles. AI
IMPACT Highlights the need for differentiated chatbot design to better support learners of varying proficiency levels in AI-scaffolded writing environments.
RANK_REASON The cluster contains an academic paper detailing a study on AI chatbot interactions.
- arXiv
- Chat Loop
- generative artificial intelligence
- Japanese
- Penny
- Revision Loop
- Transition Network Analysis
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →