A new app which aims to be “the missing piece of the puzzle for language learners” has been launched by London-based startup, Kaizen Languages, in a bid to provide low-cost, AI tutors for learners to practice written and spoken vocabulary within the context of a conversation.
Currently available for those wanting to study Japanese, the inspiration for the app came from Kaizen chief executive and co-founder Stuart Barrass’ own experiences learning Japanese and Mandarin.
“The AI will be able to identify what types of mistakes you make”
“One of the problems that most language learners face is having feedback on their speaking ability,” he said.
“If I made a mistake, sometimes people don’t correct you because it can be awkward, or you can actually make a mistake and have no instant way of knowing you made a mistake.”
Barrass added his company’s approach was to provide learners with an environment free of classroom pressures.
“One of the great powers of AI conversation is AI will correct you and give you feedback instantly; you can practice with AI without shame or embarrassment if you get something wrong,” he said.
Speaking with The PIE News, Barrass said Kaizen’s approach also provided conversational context to help learners with correct word usage, such as how a native speaker would talk with a friend compared to an authoritative figure, such as a teacher or manager.
To keep learners motivated, Kaizen also includes a points system and a streaks feature which counts the number of consecutive days a user logs on, with an achievements element planned for release.
While the app is teaching users, Barrass said it was also learning from users to replicate the experience of a one-on-one tutor.
“The more users talk to the AI, the AI is learning about the user as well, as we ultimately get to a position where the AI will be able to identify what types of mistakes you make,” he said.
Kaizen is currently developing language curriculums for Chinese, German, English, Spanish and French, and an Android app will soon be available.
