EnglishBot: An AI-Powered Conversational System for Second Language Learning
summary
EnglishBot is a voice-based chatbot with deep-learning speech recognition (94.5% speech recognition accuracy), adaptive English feedback and Chinese hints, tested with 56 Chinese college students across two six-day studies comparing conversational versus listen-and-repeat UIs. Under fixed daily assignments it matches traditional systems on IELTS-style scores but requires 2.5x practice time; under free usage it raises engagement, users spend 2.1x time spent and show significant script-based fluency gains. Results claim conversational UIs help casual self-study but are less efficient in rigid classes.
AI-written summary · about
chapters (6)
related talks
18mChatbots and the future of online learning
27mA primer to write a conversational chatbot
19mBigBlueBot: teaching strategies for successful human-agent interactions
01h 04mBuilding Expertise Into Generated Conversations
32mChatBot Learning Assistant with Large Language Models
35mDesigning Experiences for Chatbots
50mThe Art of Coding a Conversation: Designing Bots
10mBuilding Conversational AI Agents at Scale