What's the deal with using AI in replies in topics?

[CptCharis]

Hello my good friend, good to hear from you.

If you have not visited for a few months, be sure you catch up with a Rule addition regarding AI, which Rob announced here

https://www.linux.org/threads/terms...ance-on-ai-generated-posts-and-replies.58834/

...and is now listed in the Terms and Rules

Fair winds and safe seas.

Chris

Dear @wizardfromoz, I am very glad to hear from you.

Grabbing the opportunity I would like, to express my confusion regarding the negativity toward AI assistance in the forum. For people like me, for whom English is not the native language and writing isn't our strongest skill, AI has been a tremendous help. It enables us to understand complex topics, contribute meaningfully to discussions, and find accurate answers to the issues members post.

At the end of the day, AI is simply a fast and intelligent tool—like a highly advanced browser—that helps us access and process information more efficiently. It's not about replacing human input but enhancing our ability to participate and learn.

I hope this perspective can be considered as we continue to build a supportive and inclusive community.

Best regards,

THE ABOVE MESSAGE IS 100% AI

 

[kibasnowpaw]

Dear @wizardfromoz, I am very glad to hear from you.

Grabbing the opportunity I would like, to express my confusion regarding the negativity toward AI assistance in the forum. For people like me, for whom English is not the native language and writing isn't our strongest skill, AI has been a tremendous help. It enables us to understand complex topics, contribute meaningfully to discussions, and find accurate answers to the issues members post.

At the end of the day, AI is simply a fast and intelligent tool—like a highly advanced browser—that helps us access and process information more efficiently. It's not about replacing human input but enhancing our ability to participate and learn.

I hope this perspective can be considered as we continue to build a supportive and inclusive community.

Best regards,

THE ABOVE MESSAGE IS 100% AI

AI is fine on this forum you just need to give credit where it’s due. You can do it like I do and put it in your signature so there’s no doubt that you’re using it, or clearly state it each time you post. I think that’s the most important part. I use it myself.

 

[CptCharis]

AI is fine on this forum you just need to give credit where it’s due. You can do it like I do and put it in your signature so there’s no doubt that you’re using it, or clearly state it each time you post. I think that’s the most important part. I use it myself.

I really not get it, sorry. Why I have to give credits to a robot? Anyway, rules are rules

 

[f33dm3bits]

Grabbing the opportunity I would like, to express my confusion regarding the negativity toward AI assistance in the forum. For people like me, for whom English is not the native language and writing isn't our strongest skill, AI has been a tremendous help. It enables us to understand complex topics, contribute meaningfully to discussions, and find accurate answers to the issues members post.

We don't want the forum to become infested with AI posts like Youtube has been infested with AI videos, people come here for the human interaction. Getting answered as though your topic replies are only written by AI chat bots each time can be off putting to people. Yes tools are good to use, but it's also good not to become solely reliable of them to the point that you can't communicate effectively without them. There are more people here who's first language isn't English, who rarely or never use AI for replies and there are even native English speakers who use AI for some replies but not all. It's a matter of good balance, so the forum is still contains mostly human interaction and not mostly chat bot interaction.

P.S My first language isn't English either.

 

[kibasnowpaw]

I really not get it, sorry. Why I have to give credits to a robot? Anyway, rules are rules

It’s not really about “giving credit to a robot” it’s more about transparency for the people reading.

AI can sound very confident and correct even when it’s wrong or missing context. If someone doesn’t know that part of the text comes from AI, they might assume everything is fully verified or based purely on personal experience. So mentioning AI use just helps others read it with the right mindset and maybe take it with a small grain of salt.

Also, when AI helps structure or generate the wording, it’s basically a tool doing part of the heavy lifting. By saying you used AI, you’re simply clarifying that you used a tool to help express the idea just like someone might mention using translation software or other assistance.

For me personally, it’s not about who “owns” the words it’s about honesty with the community so everyone knows how the post was created.

also what he said

We don't want the forum to become infested with AI posts like Youtube has been infested with AI videos, people come here and the human interaction. Getting answered as though your topic replies are only written by I chat bots only each time can be off putting to people. Yes tools are good to use, but it's also good not to become solely reliable of them to the point that you can communicate effectively without them. There are more people here who's first language isn't English, who rarely or never use AI for replies and there are even native English speakers who use AI for some replies but not all. It's a matter of good balance, so the forum is still contains mostly human interaction and not mostly chat bot interaction.

 

[f33dm3bits]

I moved this side topic from the other topic into a new topic, so it can continue being discussed.

 

[dos2unix]

There is a "some" ethics involved here, but I don't have it all quite figured out yet.
I recently used Ai, to explian the difference between file types, if you think I didn't know this before I read the Ai port,
I have some news for you. But having Ai type it up for me, correct my speeking and grammar was handy. I was in
a hurry. I have tried not to do this. I have been typing up my articles by hand for the most part lately.

Sometimes Ai is wrong!! I always scan through the answers it gives before I post them.

Another problem... I have writeen quite a few "how to's, on quote few forums...
So now Ai, uses me as a source. ('m not the only one). In fact, Ai knows nothing it gets all of it's info
from somewhere/somebody else.. (including me apparently).. I asked for a source, and sometimes I am the source.

So... there's the ethical $$$ problem. Kind of like having one redhat subscription to get answers for my other 200
Rocky servers. The industry knows people do this, so it charges accordingly.

If I pay, say $30.00 a month for Ai, to answer other people's questions, am I robbing the Ai company of revenue?
On the other hand, if they are using my material as a source, to give other peole answers, do they owe me money?
All the people paying the Ai company to get an answer that originated from me.. should I get a commission?

I hate to break this to you, but I use Ai much more for non-Linux things, than Linux related content.
(How do I replace the headlight fuse in my truck,was one I did earlier today).

..and even Ai is opinionated, biased, and leans certian ways, It even admits it when pressed.
(which Linux distro do you like?)

Like many others, it isn't I don't know how to do something... if fact, much of my Ai prompting is ..
"put this command in", "show these flags", explain, this part of the systemd file", etc... so most of the content
is still coming from me.

I don have a problem giving Ai credit, but there is some concern about "admitting it". Usually people can tell just from the format
of the text, but now if I say I used Ai, is the Ai company going to come back and say "you shared this info we charged you for
with the whole world for free".

 

[dos2unix]

I ran my original post above thru Ai (it was completely 100% typed ny me, no Ai at all)
Here is the Ai "cleaned up" version ran thru Ai.

There is some "ethics" involved here that I’m still trying to wrap my head around.
I recently used AI to explain the difference between certain file types. To be clear: I already knew the answer. But having the AI type it up, correct my grammar, and polish the phrasing was handy because I was in a hurry. Lately, I’ve tried to avoid this and type my articles by hand, but the "convenience factor" is real.
The Accuracy Problem
The problem is that AI is frequently wrong. I’ve seen it hallucinate things that are flat-out impossible—like confirming a "Star Wars" scene where Spock kills Romulans just because I prompted it that way. It’s a "yes-man." This is why I always scan the answers before I post. If you don't have the expertise to spot the lie, the AI is a liability.
The "Circular" Ethical Problem
I’ve written a lot of "how-to" guides on various forums over the years. Now, when I ask an AI for a source, it sometimes points back to me. AI doesn't "know" anything; it’s just a mirror reflecting the info it got from somewhere else—including my own work.
This creates a weird financial loop:

1. I pay a subscription fee to the AI company.
2. The AI uses my own data to answer questions.
3. It provides those answers to other paying users.

Am I robbing the AI company of revenue by sharing their output? Or do they owe me a commission because my expertise is the "fuel" for their engine? It's a lot like having one Red Hat subscription to manage 200 Rocky servers—the industry knows it's happening, but the ethics are messy.
The Reality of Use
I use AI more for non-Linux things (like finding a headlight fuse for my truck) than for technical work. Even then, the AI is opinionated and biased based on its training data. It isn't that I don't know how to do the work; I’m usually providing the commands and the flags myself and just asking the AI to "explain this systemd file."
I don't mind giving the AI credit, but I worry about the "admission." If I admit I used it, will the AI company claim I'm giving away "their" product for free? Or will the community think I don't know my stuff?

 

[KGIII]

I've mentioned the reasons before, but I'll elaborate a tiny bit.

First, claiming the output of "AI" as your own work (by implication or not disclosing it) is akin to academic misconduct -- specifically plagiarism.

Secondly, in the academic world, you provide citations where citations are due. That is basically saying, "I got this information from this source, which I believe is a reputable source that has also done due diligence".

So, pointing out that AI was used is just a good thing.

What needs to change is the stigma some people place on the work. If the output is correct, it is a great tool. (It's also often correct. I find very few inaccuracies in the limited exposure I have at search engines.)

There is nothing wrong with using AI. It's a tool.

Using the tool analogy a bit...

We all assume that your work was 'handmade'.
If it were AI-assisted, you should make that known. We default to the assumption that it was 'handmade'.

In the physical world, if you make a product by hand, you can call it handmade.
If you make a product with a CNC machine (or 3D printer), you wouldn't call it handmade.

If the user does this regularly, they only need to add a line to their signature. If the user doesn't do this regularly, they can just add a quick blurb to tell us when they do use it.

These aren't onerous tasks. This is not a draconian rule. It's just asking for intellectual honesty, which is a good quality for a person to have.

 

[wizardfromoz]

Capta, gee I'm glad you posted this when I was asleep, lol. By the time I came to it, my friends and fellow Staff have already provided enough input that I hardly need to say a thing that has not already been covered.

I will add this -

I suggested to @kibasnowpaw the other day, elsewhere

@kibasnowpaw , with regard to that last line in your Post above -

If you edit your signature, where its first line (the same notification) appears to

• Change the font colour for that line to blue, and
• Increase the font size for that line to match that of the text you print in the Post

That will remove your need to put the disclaimer in the Post itself.

Blue font colour is quite visible on this site whether the user has Light Mode or Dark Mode on.

HTH

Wizard

I see above that he has done that, and it stands out very well.

If you only want it to appear when you are using AI-generated material to help users, to qualify what David @KGIII has said above, you could write a short snippet of text, assign that to a keyboard combination, and simply tap to insert when needed.

It might be as simple as, for example,

"Parts of the above written with help from ChatGPT"

...and the condition is satisfied.

HTH

Chris

 

[kibasnowpaw]

@dos2unix

I actually think you bring up interesting points, but I don’t fully agree with the ethical concern the way you frame it.

First thing people learning from other people’s information has always existed. Forums, books, articles, tutorials, conversations none of that suddenly becomes morally different just because AI is involved. If I read something you wrote for free on a forum and later use that knowledge to help someone else, do I owe you money? No. Because you don’t own the information itself you own your expression and wording, your style, your exact content, but not the raw knowledge behind it.

That’s the part I think gets misunderstood. AI learning from data is not fundamentally different from humans learning from data. We all absorb information from others and then re-express it in our own words. If we start saying that learning from publicly shared knowledge creates ownership rights over future explanations, then honestly the entire internet breaks.

So the “circular ethical problem” you describe where AI learns from you and then gives answers to others is not really new. People have been quoting, summarizing, and teaching each other’s knowledge forever.


About the money side:

If you pay for AI and then share the results publicly, I don’t see that as robbing the AI company. You’re paying for access to the tool, not for exclusive ownership of the output. It’s like paying for an IDE or a Linux subscription they don’t own what you create with it.

And the reverse question does the AI company owe you money because it learned from your posts? That also doesn’t really hold, because again, knowledge itself isn’t owned. Otherwise every programmer who learned from StackOverflow would owe someone a commission.

Now about AI being wrong here I fully agree with you. AI can sound extremely confident while being wrong, and that’s exactly why transparency matters. Not because we need to “credit a robot,” but because readers need context. If they know AI helped structure something, they can read it differently and double-check details if needed.

The part where you worry about “admitting” AI use honestly sounds more like social pressure than a real legal risk. The company isn’t going to come after you for sharing text you generated that would break their own business model. The real issue is community trust, not ownership.

And I also agree with you that most experienced users aren’t asking AI because they don’t know anything they use it to speed up writing, formatting, explaining, or organizing thoughts. That’s exactly how I use it too.

And honestly, I think some of the fear around AI comes from treating it like something fundamentally different from normal learning tools, when in reality it’s just a faster way to gather and organize information.


Just my viewpoint .

 

[KGIII]

If you pay for AI and then share the results publicly, I don’t see that as robbing the AI company. You’re paying for access to the tool, not for exclusive ownership of the output. It’s like paying for an IDE or a Linux subscription they don’t own what you create with it.

I'll remark on this.

If you're paying to access a scientific journal (which is both the norm and very expensive), you're still obligated to cite the papers you're using to support your own research.

Just my viewpoint .

If nothing else, it is not onerous - nor draconian.

As an aside, Draco was an Athenian lawmaker who made a bunch of laws, often making death the punishment. This was their first legal code, and it even included a council. The happy ending is that he died, and they repealed almost all of the laws not long after. I forgot the era, and Google is so far away...

 

[kibasnowpaw]

I'll remark on this.

If you're paying to access a scientific journal (which is both the norm and very expensive), you're still obligated to cite the papers you're using to support your own research.



If nothing else, it is not onerous - nor draconian.

As an aside, Draco was an Athenian lawmaker who made a bunch of laws, often making death the punishment. This was their first legal code, and it even included a council. The happy ending is that he died, and they repealed almost all of the laws not long after. I forgot the era, and Google is so far away...

Yeah, that’s a fair comparison but I still think there’s an important difference

With scientific journals, citations are required because you’re directly building on specific sources and claims, and the whole academic system depends on traceability and verification.

AI feels different to me. Yes, it learns from large amounts of publicly available information (and possibly licensed datasets), but it doesn’t give you a specific paper or exact source it generates a response based on patterns across many sources combined. So it’s not always acting like a single citation-worthy reference. Sometimes it helps structure or accelerate writing, but other times it’s also synthesizing general knowledge it has learned from many places rather than pointing to one identifiable origin.

That’s why I personally see transparency about AI use as more about context and honesty with the reader, not academic-style citation. Readers should know AI helped because AI can be wrong or overly confident, so they should take it with a grain of salt and double-check things if needed.

And honestly, people learning from shared information and then re-explaining it has always been how forums and communities work. The presence of AI doesn’t fundamentally change that it just speeds it up.

Also thanks for the Draco history side note now I’m imagining forum rules where forgetting a citation means instant exile. Probably good thing that system didn’t survive.

 

[ron.alan]

As someone who has asked for help plenty of times here and elsewhere, I want to know if someone who's trying to help me got their answer from AI.

I don't think that is unreasonable.

 

[kibasnowpaw]

As someone who has asked for help plenty of times here and elsewhere, I want to know if someone who's trying to help me got their answer from AI.

I don't think that is unreasonable.

idk if I ever have, but if you know I have, then yes there’s a 90% chance I got it from AI if it was in the past 4–5 years. And no, it’s not unreasonable. And I need to go to bed because I completely read you comment wrong anyway >_<

 

[KGIII]

With scientific journals, citations are required because you’re directly building on specific sources and claims, and the whole academic system depends on traceability and verification.

I'd like to think we hold many of those ideals as true here on this site. While we don't require as much rigor, it's still a laudable goal.

You'll note that I sometimes respond with a link to content. I could copy that and share it as though it was me writing the work, but I'd rather they go to the source.

There's also the plagiarism aspect. You're presenting information as though it came from you, when you really just copied the work from some other source.

Yes, it learns from large amounts of publicly available information (and possibly licensed datasets), but it doesn’t give you a specific paper or exact source it generates a response based on patterns across many sources combined.

Absolutely, but it also gives people a starting point if they wish to dig deeper, verify data, etc..

 

[kibasnowpaw]

I'd like to think we hold many of those ideals as true here on this site. While we don't require as much rigor, it's still a laudable goal.

You'll note that I sometimes respond with a link to content. I could copy that and share it as though it was me writing the work, but I'd rather they go to the source.

There's also the plagiarism aspect. You're presenting information as though it came from you, when you really just copied the work from some other source.



Absolutely, but it also gives people a starting point if they wish to dig deeper, verify data, etc..

Yeah, on the plagiarism part I actually completely agree with you.

And honestly, I’ll admit something it’s not always something I actively think about in the moment. My main focus when I reply is usually just trying to help the person and trying to be as correct and factual as possible while still keeping my own opinion and viewpoint intact.

The way I personally use AI is more like this:

I give it my thoughts, my opinions, what I want to say, and the angle I’m coming from. Then I ask it to help structure it, clean up grammar, and double-check factual accuracy. So the end result is usually a mix my ideas and experience, but with AI helping organize and refine the wording.

After that I normally run it through TTS and listen to it. If it sounds like something I actually agree with, I post it. If not, I go back, tell it what feels wrong or off, and adjust it until it reflects what I actually mean.

I also try not to go deep into things I don’t personally understand or have experience with, because I don’t have the real-world knowledge to back that up. Sometimes AI suggests commands or tools I haven’t used myself and I try to avoid relying on those unless I think they genuinely help and make sense, but yeah, it can happen occasionally.

So for me it’s less about copying someone else’s work and more about using a tool to help communicate what I already want to say but I do understand your point about attribution and transparency, and I think that’s a fair concern.

 

[KGIII]

I give it my thoughts, my opinions, what I want to say, and the angle I’m coming from. Then I ask it to help structure it, clean up grammar, and double-check factual accuracy. So the end result is usually a mix my ideas and experience, but with AI helping organize and refine the wording.

My only real use of AI is the blurb at the top of Google. (I have played with AI in the past, even using it to create music.)

I've found my search practices have changed. With the current situation, you can ask complicated questions and get an answer that's not a direct search result. (You can verify the sources.) It'll even do some fairly complicated mathematics and explain complicated science. So, my searching has changed to take advantage of that.