Practicing Japanese with ChatGPT

Hello everyone,

You probably already heard about ChatGPT so I won’t get into too much detail about it,
in a nutshell, ChatGPT is an AI language model that interacts in a conversational way, making it seem human like in his responses, however, people have used it for things that are far beyond that, like asking for an entire sales pitch for a product or writing actual working lines of code, which ChatGPT managed to deliver upon successfully.

Like many people using WaniKani, I encountered the problem of practical application of what I learned here, since learning a language in a vacuum and actually using it are two completely different things.
and like many other people here, I tried applying what I learned so far (I’m currently at level 35) in various ways, like reading books/manga/light novel, playing games fully in Japanese, etc., but one thing that is still missing for me is writing in Japanese, writing actual sentences and not just following a question in a book or translating an English word to it’s Kanji form.

So I tried using ChatGPT in a more ‘practical’ fashion, like practicing a new language for example, and so far it seems like a promising addition to the repertoire of tools to use for practicing Japanese.
It can understand my writing, even when it’s somewhat broken and grammatically inaccurate, and provide interesting and detailed answers, which is great for reading practice as well.

So far I really enjoy using ChatGPT this way, and from a learning and practicing perspective, enjoyment is a big factor for success and perseverance.
I was wondering if other people tried using ChatGPT in this way and what their results were.

I haven’t used it but there was a post about it quite recently. It can and will make mistakes, while being confident about it. Honestly, I wouldn’t recommend it.


The problem is it will very confidently tell you things that are completely wrong. This recent thread has some good examples.


I’m no expert but, while humans aren’t perfect, a good teacher will spot and almost all of your mistakes and won’t reply back with something excessively incorrect, AIs are just not at that level. While they ‘understand’ you, they won’t give you feedback why you are wrong.

You’re risking entering an echo chamber where your mistakes are constanly being reinforced, which is even worse for your studying.