Have any of you tried using ChatGPT for simple questions? How reliable is it?

An example of a recent conversation I had with ChatGPT:

> **Me:** Hello! What is the difference between 対面 and 面会??
>
> **ChatGPT:** Both 対面 and 面会 are Japanese terms that refer to a meeting or an encounter with someone in person. However, there is a subtle difference between the two.
>
> 対面 (taimen) refers to a face-to-face meeting or encounter, which can be formal or informal. It implies that two people are meeting in person and looking directly at each other. This term is often used in a business or professional context.
>
> 面会 (menkai) also means a face-to-face meeting, but it is more commonly used in a legal or correctional context. It refers specifically to a meeting with someone who is in custody or imprisoned, such as a prisoner or a defendant. In this context, the term implies a more formal meeting, often with restrictions or supervision.
>
> In summary, while both terms refer to in-person meetings, 対面 is a more general term that can be used in various contexts, whereas 面会 is a more specific term often used in a legal or correctional context.

Is the AI correct?

Have any of you tried asking chatGPT questions and received false information? I wonder if I can trust its replies.

4 comments
  1. >Have any of you tried asking chatGPT questions and received false information? I wonder if I can trust its replies.

    Don’t trust it. What it delivers ranges from correct, to half correct and half wrong, to complet bs. As a beginner or even intermediat learner it is hard to recognize when it is wrong. Even as an advanced learner it can make you second guess, as it writes with so much confidence.If you have to check whether its correct or not anyway, you can just do so right away, bypassing GPT.

    PS: Search for GPT in the subreddit. There have been many questions similar to yours, and the answers are always the same.

  2. Never let machines make your mistakes for you. If you can’t answer the question in the post title yourself, you should consider AI wrong all the time. It’s not, but you might as well. If you **do** have enough knowledge to discern that, there’s a good chance you’d also have the experience to keep up with native explanations from… well… NATIVES who actually use the language. In which case, the AI would be worthless.

  3. ChatGPT has been really helpful for me. Sometimes, when trying to find the meaning of something that is conjugated, I run into difficulties as I dont know which word it is. I asked chatgpt to split the sentence up and so far it’s been right. Of course, I double-check afterward.
    I do this because Jisho doesn’t always help with these things.
    It’s definitely more useful than people here make it out to be. I wouldn’t avoid it like the plague, but you do need to keep in mind that it will be wrong. Use it as a resource, but take the answers with a grain of salt.

  4. In your specific example, I would say it failed to mention other ways those two words are differentiated (in how they’re used).

    Consulting a [monolingual dictionary](https://dictionary.goo.ne.jp/word/%E9%9D%A2%E4%BC%9A/), there’s a bunch of factors that differentiate the usage of 面会, 面接, and 対面. The points described here are mostly not mentioned at all in the response by ChatGPT.

    So… you still risk missing out on important information by using ChatGPT even if it sounds good, i.e. you practically need to double check every damn time you ask it something. I’d just say google your question and cut out the initial ChatGPT, because most of the time, you’re not the first one to ask.

    Similar to what I said on a different post, the “person who makes an error but doesn’t realize that they did” will obviously say “they don’t make an error”. To say “I don’t make any error(s)”, there needs to be a knowledgeable, experienced auditor (e.g. tutors) to go over their understanding and whatnot. Can’t count how much mistakes that they overlooked, obviously.
    Which, admittedly, makes it a judgment call to decide which is better, the “don’t use it, it makes errors and you’ll need to double check anyways while risking to consume crap” or the “use in moderation while cross-checking the accuracy”. Your choice.

Leave a Reply
You May Also Like