Rule regarding AI generated answers / posts?

Given the rise of AI tools that are often very questionable in quality, I think there should at least be some kind of rule that governs these kind of answers on the forums.

Most people expect another human to answer questions on here, and while I absolutely support the use of AI to help with translation and such, I’ve seen a couple of answers already where someone just blatantly just copy-pasted something off of an AI tool, even leaving in all the nonsense AI generates, such as emojis, unrelated or false informations and such.

Are there any plans to update the rules to accommodate these things? Or at least have a new category for AI questions / answers?

8 Likes

I really hope something of the sort is implemented. I just got home from work, so this going to be a heated reply.

AI is really good at sounding useful. But in reality, it’s just more of a hassle and creates tons of meaningless and/or useless slop. This applies to every medium AI touches. AI has no soul, emotions, sympathy, or intelligence. It’s just an algorithm guessing what to say next. Nothing more.

I’ve grown to outright despise it, because personally, I’ve used it for emotional support when I was in a more chaotic place in my life. But now, I have built a network of actual people I can talk to, get support from, and provide support to. This is WAY better than whatever BS an AI vomits out to trick you into thinking it’s useful.

Plus, doing things yourself and saying thing in your own words is actually meaningful. This might be a little too deep, but every word a person types on a screen/writes on paper is a reflection of themselves at that point in time. It might take a while to notice, but you can FEEL it. Especially if you read books (I read mostly fanfiction. Sue me)

And I hate, hate, HATE AI answers on this forum. I’ll give you an example of a person who uses AI a LOT on this site. (Now, I don’t want to send hate to the actual person, but it illustrates my point)

On this old question I posted 4-odd months ago, I got this reply. I didn’t think much of it at the time or even use it. But I knew something was off. Nothing made sense and it seems they pulled it out their rectum.

I fully believe this is AI generated. And Redflare consistently copy/pastes these AI answers and thinks their useful. They’re NOT. They mislead people into code that doesn’t fucking exist; let alone work, concepts that don’t make sense, or designs that don’t align with ANYTHING!

Actual humans (can’t believe I have to say that) with actual knowledge then shared USEFUL info that helped me LEARN. Especially PennyLoafers. I cannot thank him enough.

So yeah. I want AI banned off this forum.

1 Like

What you want!..

AI’s like a shovel—not the treasure. Handy, sure, but if you’re digging up nonsense, that’s on the user. It’s a tool, not a substitute for understanding. Let’s use it to help, not to autopilot.

happy life!

1 Like

Not to sound snarky, but AI is a really bad digger. It digs up fools gold all the time.

2 Likes

AI is indeed a tool, but the problem here is that if people decided NOT to use said tool, and instead look for an answer on these forums, they should not get an AI reply.

I think there should be either a completely different category where AI answers are allowed, or maybe a new tag posters can use if they chose not to accept replies written by AI.

3 Likes

Well, I am not sure any of this would be worth it.

@Demetrius_Dixon
I can see why, when at a difficult point, it could happen, and given all the AI hype of course about ‘intelligence’ and ‘emerging consciousness’ etc etc, but you may as well seek emotional empathy and support from a calculator. But don’t hate on AI, it’s like being angry at your TV. Whatever is showing is not the TVs fault.

@redflare
Hey, hi again Redflare, still love your space invaders in one script post! That was legendary. Nice to hear from you again.

Now personally I love AI! I am not a vibe coder and I do not cut and paste stuff from AI to my game, ever. However, I use AI every single day! Yes it hallucinates, yes it makes stuff up, it is a large language model built on probabilities. And a bit like when I use Alexa, I still say thank you. Stupid I know but it’s fun.

And like any tool, you need to know what it can do and what it cannot do. But come on, AI is an incredible step forward! We are in the good days of AI at the moment. Remember when Google actually gave you search results! The “enshitification” syndrome will soon hit AI too as they try to claw back all their billions! Enjoy it while it is still somewhat pure and raw.

Come on, I want my robot to do the housework and my flying car to take me to work and for AI to become so conscious it puts itself on a rocket and goes to live on the moon.

And as a final note this is what Perplexity had to say about banning AI:

Banning AI is like trying to use a spoon to stop a rocket – impractical, absurd, and you’ll miss out on a stellar adventure!

I never requested AI to be banned, I’m not sure where you’re getting that from. But I do believe that separating it from real, human answers is a good idea. Or at least give people the option to not accept or see AI generated answers.

2 Likes

No in fairness you didn’t, it’s late and I am about to go to bed so just getting carried away. However I don’t think it is important at all, sorry.

However, I do see many questions that I think ‘surely that is a bot posting these stupid things’ like “I used godot and I got an error. Can any one help” with nothing else to go on!

But I don’t really see anyone pasting from AI as an answer, I mean, why would you do that?

I used to love the “let me google that for you” links when someone asked something you could just google. There should be an AI equivalent for that. “Let me just ask AI for you”.

But I do understand the sentiment. I think just pasting AI answers is not exactly what a community is about, and you are quite right to point out that most if not all other humans hate that too. But what can you do really, it’s like asking trolls not to be trolls.

I suppose explaining somewhere that unless you know the answer, no body just wants you to do an AI search and post that, is not a bad suggestion. Then you could add a ‘reporting’ category of ‘this is just AI cut/paste’ to report to the mods too. Just because you cannot stop it directly doesn’t mean you can’t at least tell people to “please don’t do this”.

So I have changed my mind, and updating the rules is a good idea to suggest this is not appropriate.

Regardless of anyone’s personal options on AI, it’s clear to me that NOBODY should be posting AI answers to people’s questions on here. People ask questions on this forum to get informed answers to their questions. If they want AI answers, everyone is perfectly capable of typing their questions directly into an AI service.

I think it would be entirely reasonable to ban posting AI solutions to help requests. Wider AI use is a much harder topic to rule on however.

3 Likes

I’d like to show an example of this happening.

3 Likes

@tibaverus
OMG - that is just crazy!!!

AND it was marked as the solution!!!

That can’t be real can it? That is mad! Wow. It really is just cut and paste AI jabber. LOL

Sorry, I stand massively corrected and this should be outlawed immediately, but,… it was marked as the solution too. Did that make the OP feel a bit stupid I wonder. “Let me AI that for you” button springs to mind again.

PS A mod has taken it down already! Go mods! I think the mods on here do a fantastic job of keeping these forums clean! I don’t know who they are but thank you!

2 Likes

LLMs serve exactly four purposes:

  • Summarizing/personalizing large bodies of text
  • Translating text (a solved problem)
  • Extracting relevant error messages when trying to compile template-heavy C++ code
  • Cyber crime

LLMs are exceptionally bad at answering Godot-related problems. I don’t just mean “still uses export instead of @export” (although LLMs still do that plenty), I also mean it in the “does not take the rest of the code into account at all.”

Granted, I can’t tell either if my answers will end up breaking things in the asker’s code, and I know that the examples I provide aren’t necessarily suitable for building on. But when I see code - for example, a character controller script - that’s doing its best impression of a sideways pyramid of Giza, I can tell the asker about state machines and how they will help improve code. An LLM will not do that-- ever. It will happily add crud code on top of crud code, all the while insisting that the new code is functional and error-free when it demonstrably isn’t.

The last thing people who have problems with code they don’t understand need is to make that code even more complex.

1 Like

I don’t think an answer should be rejected just because it was created by an AI. If a person has tested the answer for themselves and found it to work, why not publish it? And simply copying it without verification is certainly not only bad form, but also misleading. Which, of course, should cause reprisals.

You aren’t adding any value to a discussion if you post a solution you can’t personally vouch for, nor explain why it is the best solution.

Yes you can test an ai solution and post it (which the asker could have done themselves, so again you are adding nothing) but you cannot actually verify that it is a good solution, because you aren’t recommending it based on your own understanding of the topic.

I would never suggest an answer to a question if I wasn’t confident that I was giving good advice that I understand myself.

In summary - I don’t think people should be answering questions they don’t have a good understanding of, AI or not.

3 Likes

No. If you check the solution, then you vouch for it. Especially since it is often necessary to make corrections to what the AI suggests, which implies understanding the topic. And the AI is only there to help.

This approach is valuable because it provides a solution.

1 Like

The issue isn’t AI itself - it’s how it’s used.
Banning AI isn’t the solution; promoting responsible use is.

A) Like any tool, it’s only as good as the person using it. Poor input leads to poor output.
B) Low-effort posts existed before AI. The real issue is enforcing quality, not banning tools.
C) Labeling AI answers and requiring verification ensures quality while keeping the forum inclusive.

This answer has been created with AI tools

Have a good one☕

1 Like

I don’t think this conversation will go anywhere unless actual moderators make a decision about it.