Like full disclosure I absolutely use AI to check my keigo when I message colleagues so it’s not like im fully against it, I just don’t think it’s necessary to be used on this site (tho there are times im pretty against it tbf)
Already lots of sites are adding it so much that it’s making people upset, so from a business perspective I think some people would be upset if they felt bunpro was bandwagoning basically
Glad to hear that you had such a moment!
Every time I try to use it to correct a comment I want to leave under a YouTube video (like today) I then start burring myself under a bunch of thoughts like “Oh now I’ll never master this language, it’s all pointless”
No frrr I wish it didn’t always jump straight to correcting, it’s deeply frustrating. But that’s the thing tho I think it corrects even things that are perfect, so I try to always preface with like “only tell me if something is really weird” or something like that, but even then the doubt sets in and it’s easy to lose confidence in your conversational abilities
Prompt engineering is a difficult thing to get right but it’s vital if you wanna use AI for language learning. Because of that I think a lot of people have had really bad experiences because of just not knowing some good methods of using AI as a tool. This is a really good video that can probably help with solving a lot of your confusion and hopefully help with letting you see how well AI can work if you plan on going at it from just a text input standpoint.
How to engineer good prompts: https://www.youtube.com/watch?v=p09yRj47kNM
I feel like a broken record here but there are more ways to use AI than just a question in and response out chatbot and I really encourage people to look into and learn about the ways that it’s being used. You can even see quite a very examples that others have said above that aren’t just a chatbot.
The biggest issue I see is most people using AI just aren’t good at it and it can lead to some pretty negative results which I have seen myself and others have referenced above. If you still want to use it or maybe get a better understanding of why some people like AI and the ways it can be used then I really recommend looking into it more. Also just one last time AI isn’t just chatbots that you talk to!
I think the biggest thing is that no matter what you say the trust just isn’t there in the system like
And that’s what makes it a bad move 
Business isn’t just about facts but about responding to user/market feelings and perceptions
I’m a professor and I absolutely despise generative AI. Academia was already a sinking ship and generative AI is like a shotgun to the hull.
The way the human brain works, it’ll grasp at the first shortcut it can find. I’ve watched my students and colleagues give up on anything challenging sooner and sooner because the magic word-guesser will do it for them. Everyone thinks they’re immune to that pull and everyone is wrong. If you let something do your work for you, your ability to do it yourself will atrophy. It’s human nature.
Besides, I oppose generative AI in Bunpro on philosophical grounds: languages exist for human connection. Which is in short enough supply as it is.
Personally, the closest I come is using DeepL. I’m not proud of it, but even then, I only use it after getting as far as I can without it. And I stop and check its translation against the original, particle by particle, every time. As soon as you lose the trying-it-yourself step, you lose the learning.
I just don’t see a way to work in generative AI that wouldn’t either compromise learning or compromise human connection.
Bad news: LLMs also use stolen material. It got those grammar explanations from people that wrote them.
Also: to the other folks that are trying to say “oh you want to discount all AI what about all these machine learning technologies then”
Let’s all get on the same page. We are talking about LLMs here. Let’s not conflate all of these. People buy the phrase ‘AI’ because LLMs are sentence writing machines and thus people think that they are reasoning machines. The popular conception of AI is directly related to the function of LLMs. No one is talking about other uses of machine learning.
bunpro got the grammar points from other sources, they didnt just make them up 
the ai model also did that. its just that human art is either completely original, or takes some inspiration from other artists instead of taking the artists art and changing it directly
holy this topic almost has 100 replies in under 24 hours
I see a dystopian future where everyone who studies a language only ends up talking to an AI in that language cause no one has a need for human connection in all our flaws anymore.
huh
Not the case for me, I’m myself a very ai person, I use it everyday, talking to it and building a prompt to get exactly what I want is what I do all the time for the time it came out, usually to me it feels like you have to have AI talking fluency, AI slang to talk to it, and understand what it say, and I have that. I can notice when it’s telling something fishy within a millisecond, when it writes code for me it always works (for the limited tasks I do know).
And know when I have to doubt it, and exactly because of that I have that feeling of doubt every time it correct me, because I know that it doesn’t understand nuance, I know that it’s specifically that context where it could make a lot of corrections without understanding how bad those mistakes would feel to a native.
No matter how I form the prompt for that, in the case of language correction that feeling always follows me when I use it. And again, in all other cases it’s super usable as for me
This is a bad take im sorry 
Edit: wait nvm I think I misunderstood your point?? I’m confused
mb 
but ai art is still not real art
No I thoguht you were defending ai art lmao im so confused
wait what
ooohh, lmao, i would never defend ai art
I’m referencing using machine learning and have several times above when talking about improving the review system and incorporating personalized learning plans based on users goals and current knowledge. I also see a use for GenAI but I understand how people can dislike it because AI needs input data to be trained on and I think that it should only be done with information people are giving up with that intention.
大丈夫！
My bad, sorry
I gotta say it’s been pretty cool to see so many people that use Bunpro show up in mass, it really makes the community feel bigger in a way!