Should chatbot be allowed in troubleshooting?

[Fendrick]

1 minute ago, Middcore said:

 

The account this thread is basically about was linked above, you could have looked at it and seen for yourself the quality of what they were contributing. The information you needed to avoid making a wrong assumption was at hand, please don't blame others for failing to spoonfeed it to you. 

I apologize.

[Middcore]

Just now, Fendrick said:

I apologize.

 

No worries.

[Mark Kaine]

1 hour ago, Fendrick said:

There will always be people who prioritize and value having knowledge and sometimes people may lack knowledge in a specific criteria as well this hobby is not a stagnant thing, using AI is not really a detriment in any way if the person it gets to is helped accurately.

 

This seems to me to be based on fear or maybe even some egotism.

nope, thats just wrong as has already been said people don't come to forums to talk to bots and thats besides that in a lot of cases a bot cant help, because it lacks nuance.

but the main reason people go onto forums is to talk to like-minded *people* that is the issue, a bot isn't  *people*, social interaction drops to zero if you talk to a bot, especially when they're hiding that fact...

 

if you want to talk to a bot you can just talk to chatbot,  you don't need a forum for that...

 

ultimately bots on forums would eventually completely destroy forums,  bots everywhere,  whats the point lol, they aren't interesting and they aren't adding anything new or worthwhile (except random facts that may be true , well, randomly) 

 

so i bet most people who care about forums don't want bots on the forums they visit, simple as that,  and its not new either,  chatbot just makes it slightly easier to hide the fact that you're "chatting" with a computer instead of a human, it has no place on a forum, its detrimental in any way and hasn't been allowed ever (for the outlined reasons...)

 

 

[Fendrick]

Just now, Mark Kaine said:

nope, thats just wrong as has already been said people don't come to forums to talk to bots and thats besides that in a lot of cases a bot cant help, because it lacks nuance.

but the main reason people go onto forums is to talk to like-minded *people* that is the issue, a bot isn't  *people*, social interaction drops to zero if you talk to a bot, especially when they're hiding that fact...

 

if you want to talk to a bot you can just talk to chatbot,  you don't need a forum for that...

 

ultimately bots on forums would eventually completely destroy forums,  bots everywhere,  whats the point lol, they aren't interesting and they aren't adding anything new or worthwhile (except random facts that may be true , well, randomly) 

 

so i bet most people who care about forums don't want bots on the forums they visit, simple as that,  and its not new either,  chatbot just makes it slightly easier to hide the fact that you're "chatting" with a computer instead of a human, it has no place on a forum, its detrimental in any way and hasn't been allowed ever (for the outlined reasons...)

 

 

Bots are bad because they are trained to be like humans, so from this we can assume being human = bad.

 

Refer to my thread. Read my reply to the posts.

 

 

 

 

[Fendrick]

Reacted in 2 seconds to my post.

No need to go forwards here.

[Mark Kaine]

2 minutes ago, Fendrick said:

Bots are bad because they are trained to be like humans, so from this we can assume being human = bad.

This is complete nonsense,  sorry.

 

Bots have no place on forums, because they're actually much worse than humans... they don't have to follow any rules either, because they have no feelings,  they simply dont care about anything and dont contribute anything worthwhile. 

[Fendrick]

3 minutes ago, Mark Kaine said:

This is complete nonsense,  sorry.

 

Bots have no place on forums, because they're actually much worse than humans... they don't have to follow any rules either, because they have no feelings,  they simply dont care about anything and dont contribute anything worthwhile. 

Because they are trained to be like humans which are imperfect.

 

The suggestions is to model AI of off rules for intellect. The idea is information, not your life story or how you feel.

 

From the other thread since you seem to have missed my proposal.

 

"I think the idea would be to train the AI of a concensus or rule guide for intellect only then appropriately add some human traits, not the other way around, humans don't need human nuances for straight to the point information, though they could come up with lists and detailed sub-information detailing the process and conclusions in the deciding information. "

 

 

Some people don't even know what intellect is and I was asked what it is so here is what you need.

 

Everything for a bot to be easily workable.

 

"Intellect as we know right now is problem solving, in an IQ test that looks for this it would be "G"

 

https://en.wikipedia.org/wiki/Problem_solving

 

https://www.i3mindware.com/5-iq-factors

 

  GQ for mathematical reasoning and problem solving.

 

Reasoning and problem solving are not the same process but are accompanying processes to what we call "G" in IQ.

 

The prefrontal cortex is the main factor in the human being, but the human beings brain is not just the prefrontal cortex, AI would be using just that and a very small amount of the rest to bring in a  human like factor.

 

https://pubmed.ncbi.nlm.nih.gov/18467667/ "

[Mark Kaine]

4 hours ago, LogicalDrm said:

Plagiarism is already within what should be reported. While not expressively said, and its mostly issue in Tech News. As academic, not properly attributing work that you didn't do is serious offense to me. If one doesn't have respect towards sources they get their information from, how can we expect them to value any rules?

we still need some kind of new report category i guess? to me "bot spam" would actually work because that's what it is, but maybe something more descriptive would be better?

[Spotty]

If you're unsure if something is against the rules of the forum either contact a forum moderator and ask or report the content and the moderators can review it and make a determination. If it's something that turns out not to be an issue we can just close the report without taking any action.

 

Using a bot to generate posts isn't expressly mentioned in the Community Standards, however using a bot to write your posts would generally fall under the No spamming or no-constructive posts. Particularly the "non constructive" aspect and post farming.

Quote

No spamming or non-constructive posts, including:

• Posting the same topic multiple times
• LMGTFY (Let Me Google That For You) links or comments in the same spirit
• Post count / Solution count / reputation farming

 

This is a tech forum so there's bound to be some interest amongst the community for chatbots and AI. It's okay if people want to have discussions about AI/chat bots or if people are using them as part of their post where doing so is relevant in the context of the discussion. As an example there's the ChatGPT AI Discussions thread where people have been posting and discussing some results from ChatGPT. Discussions such as those are fine as the discussion is related to AI/chatbots and it's clear that the responses are bot generated.

 

 

In response to the discussion on whether or not bot replies should be allowed on the forum, my own opinions are it should not be allowed...
 

Posts should be your own personal input. Using a bot to write replies is not within the spirit of a community forum.

 

When people post to the forum asking questions they don't do so expecting bot answers. Many people asking for help may be unaware that the answers they received were generated by a bot, and lacking that information be unable to determine for themselves if they want to trust the advice given by a bot or wait for a proper response. While some people might be okay with receiving a bot response others may not. This applies to not only the person who asked the question but also any other people who may read the discussion (remember your audience both present and future). Posting bot generated messages as if it was your own post is very misleading.

 

With posts that are people's personal input even if something they post is wrong that can still encourage further discussion as people correct it and people can learn from it and expand their knowledge. There's nothing constructive to be gained from trying to argue against replies that somebody just copied and pasted from a bot, and I think most people would feel that trying to correct mistakes in bot replies would just be a waste of their time. Bot replies do not encourage the same discussion that normal replies do and would turn people away from participating in discussions. This means that bot replies are inherently not constructive.

 

As for the idea of people curating bot answers before they post them; If you have the knowledge and expertise to properly curate a bot generated message to ensure its accuracy and modify the message to your own work with your own personal input then you don't need to use the chatbot to write the message for you. Generating a bot response, reviewing it, correcting and adapting it is going to be much more work than just posting what you think the answer is.

 

If left unchecked people botting replies would be an obvious issue. Nobody wants to see every post get hundreds of bot replies with chatbots talking to each other to the point where it makes any real discussion pointless or impossible. There will always need to be a line drawn somewhere.

[Spotty]

15 minutes ago, Mark Kaine said:

we still need some kind of new report category i guess? to me "bot spam" would actually work because that's what it is, but maybe something more descriptive would be better?

Please reserve the bot spam report category for advertising spam posts.

For stuff like this "Trolling or pointless content" may be appropriate, though the default "Something else" category is also fine and if you're ever in doubt which category something falls under use that. It is helpful to include a short description with reports, especially with something that might not be obvious when looking at a single post and may require additional context. Reporting it and including a comment saying "I think this person is using a bot to post replies" will give the moderators a good idea of what they should be reviewing.

[Mark Kaine]

32 minutes ago, Spotty said:

Please reserve the bot spam report category for advertising spam posts.

For stuff like this "Trolling or pointless content" may be appropriate, though the default "Something else" category is also fine and if you're ever in doubt which category something falls under use that. It is helpful to include a short description with reports, especially with something that might not be obvious when looking at a single post and may require additional context. Reporting it and including a comment saying "I think this person is using a bot to post replies" will give the moderators a good idea of what they should be reviewing.

yeah, i know,  that's what im doing usually,  i was just saying it because  i think the current report options make this a bit more complicated than it should be,  and i think a new, more descriptive option will be added eventually (doesn't have to be a big thing, like what you mentioned "pointless content" is something where it could simply be added for example, i guess)

 

40 minutes ago, Spotty said:

If left unchecked people botting replies would be an obvious issue. Nobody wants to see every post get hundreds of bot replies with chatbots talking to each other to the point where it makes any real discussion pointless or impossible. There will always need to be a line drawn somewhere.

this whole post is spot on (excuse the pun! ) especially that last paragraph is also why i think this needs to be addressed rather fast, because this whole thing has the potential to kill *any* forum real quick, unfortunately... and its not gonna be easy to prevent either, i suppose... i just think its still kinda urgent,  because it could go down quickly (doesn't have to but the potential is definitely there)

 

But I also feel / hope a forum with generally good moderation (like this one) can deal better with this stuff than others.

 

Like at first i didn't think much about it... just an odd thing... but when Mr chatbot answered in a thread I made, it instantly hit me... "this really sucks" like literally "googlefu", not helpful at all (because none of the things it said were actually relevant) and how do you even respond to something like that...? As said, talking to bots (and probably constantly correcting them on top of it) is certainly not why most people visit forums...

 

 

[Vishera]

4 hours ago, Fendrick said:

using AI is not really a detriment in any way if the person it gets to is helped accurately.

 

This seems to me to be based on fear or maybe even some egotism.

That's a big if!

But that's not the point, People come here to ask for help from us the users, if they wanted help from Google Search or ChatGPT they would have asked Google Search or ChatGPT for answers,

But no, they came here looking for answers from us!

So copy pasting answers from Google Search or ChatGPT is pretty much SPAM considering they came to the forums to ask for help from the community!

[05032-Mendicant-Bias]

In bird culture, taking the "I'm lucky" google answer or Chat GPT answer and posting it as answer to someone's question is considered a dick move. OP can do those steps, you are not adding anything. I believe AI tools are going to become ever more pervasive, so a ban feels unenforcable, tought stack overflow did just that. I think the more sensible solution is to ask for AI generated answers to be clearly labeld as such.

 

Curated AI answers are the same as curated google search. AS long as you took effort to add value to the answer, it makes no difference what tool you used to get to the answer.

[Mark Kaine]

@Visherayeah, we knew that... it's why i made this thread...

 

 

[Mark Kaine]

53 minutes ago, 05032-Mendicant-Bias said:

think the more sensible solution is to ask for AI generated answers to be clearly labeld as such.

the thing is,  that would be still highly detrimental... but it could be an option for a dedicated section of the forum for sure (theoretically)... its not like i hate spambot i mean chatbot, its the way its currently abused to basically troll, spam and farm... its super annoying and highly anti social and detrimenal to any community... so really the issue isnt chatGPT itself,  its how certain individuals use it and probably think they're being funny or clever while doing so...

 

Spoiler

(assuming its not all automated anyways) 

 

[cmndr]

This is a good question. Since I didn't know the answer, I checked with ChatGPT. 
 

Quote

It is generally a good idea for online forums to allow chatbots to be used for technical support and troubleshooting, as they can provide a quick and convenient way for users to get assistance with their issues. Chatbots can be programmed to provide users with relevant information, troubleshooting steps, and links to helpful resources, which can help users resolve their issues more efficiently.
 

However, it is important for online forums to make it clear to users that they are interacting with a chatbot rather than a human, so that users know what to expect in terms of the level of assistance and accuracy of the information provided. It is also important for the chatbot to be programmed with accurate and up-to-date information, and for the forum to have processes in place for handling more complex or unusual issues that may require human intervention.

 

That answer is good enough for me. 


 

Quote

As for the idea of people curating bot answers before they post them; If you have the knowledge and expertise to properly curate a bot generated message to ensure its accuracy and modify the message to your own work with your own personal input then you don't need to use the chatbot to write the message for you. Generating a bot response, reviewing it, correcting and adapting it is going to be much more work than just posting what you think the answer is.

You underestimate how accurate the bots are and how much time is saved in typing, along with the effort put into thinking up what to write. I DID review my chatGPT answer to ensure it was sensible. No corrections needed and it happened to mirror my own views. 

[Mark Kaine]

5 hours ago, cmndr said:

That answer is good enough for me. 

not for me ... its not a very logical answer because its not practical at all,  especially if you'd look at what this user is posting (as an example)  ... its exactly the kind of broad advice people who go on forums like this are trying to avoid (most likely) ... 

 

 

anyone can google these one size fits all answers in 5 seconds and its simply not very helpful typically and not suitable for a forum (which's main purpose is to encourage discussion,  even or especially in troubleshooting,  because of how complex things can get...)

 

5 hours ago, cmndr said:

You underestimate how accurate the bots

they aren't accurate at all,  with tech stuff... that is part of the problem,  but not the whole problem. 

 

just because they get some things right, doesn't mean they arent basically just guessing (without understanding the subject)  ive seen answers where the bot didn't even understand that pc and consoles aren't interchangeable/ compatible...

 

Like i actually think the bot is good at certain stuff, like "chatting" or stories,  or even programming... but computer tech is, ironically, the thing where it really fails, quickly and miserably. 

 

 

 

5 hours ago, cmndr said:

This is a good question. Since I didn't know the answer, I checked with ChatGPT. 

 this reads like satire... 

 

 

but, its actually a good example why the bot kinda sucks...

 

5 hours ago, cmndr said:

it is important for online forums to make it clear to users that they are interacting with a chatbot rather than a human,

it doesn't understand that this is not possible (or what a forum even is , since its missing context) because its dumb as bread... and this is exactly why it's *not* suitable for tech / support questions (with exception of really low level stuff, which im sure it would get mostly right, but again,  anyone could just google those, and typically issues are more complex) 

 

tldr: even if it could solve simple issues,  its of no use, because humans can do that too and probably better, and for higher level stuff its simply not suited because of complexity,  it would only create a lot of chaos and despair due to misinformation overload... you can't even discuss with it,  it thinks its always right, the worst poster imaginable lol

[05032-Mendicant-Bias]

10 hours ago, Mark Kaine said:

the thing is,  that would be still highly detrimental... but it could be an option for a dedicated section of the forum for sure (theoretically)... its not like i hate spambot i mean chatbot, its the way its currently abused to basically troll, spam and farm... its super annoying and highly anti social and detrimenal to any community... so really the issue isnt chatGPT itself,  its how certain individuals use it and probably think they're being funny or clever while doing so...

You hit the nail on the head.
The genie is out of the bottle, there is no putting it back. People are going to use AI chatbots. Because AI chatbots are getting good.

 

4 hours ago, Mark Kaine said:

they aren't accurate at all,  with tech stuff... that is part of the problem,  but not the whole problem. 

ChatGPT is hit and miss. When it hits it really saves a lot of time compared to a simple google search. Literally hours of work saved with one good prompt for some tasks, like finding out what framework/library to use and how to wrap it in your code.

When it misses, it's up to you to see chatgpt is bullshitting and refine the prompt or give up because that's something chatgpt can't do.

It's a tool. There are things it can do well, and things it does badly. And often allucinates wrong answers. I wouldn't ask chatgpt to choose components for my new build, I would ask here in this forum. Chatgpt hasn't tried GPUs CPUs, etc... the users have. I would guess chatgpt training data is poisoned by all the fanboy/hater discussions about products to have a balanced understanding.

 

Luke on the wan show said he is now using chatgpt first instead of google when searching stuffs. You really are underestimating how good ChatGPT it is. And it's only going to get better from here on.

[Vishera]

41 minutes ago, 05032-Mendicant-Bias said:

Luke on the wan show said he is now using chatgpt first instead of google when searching stuffs. You really are underestimating how good ChatGPT it is. And it's only going to get better from here on.

Considering how machine learning works you probably can teach ChatGPT to be stupid, especially since bots don't have the capacity to understand.

They may have the ability to learn with machine learning, but bots don't have the capacity to understand - They just do what they are programmed to do.

[LogicalDrm]

3 hours ago, 05032-Mendicant-Bias said:

You hit the nail on the head.
The genie is out of the bottle, there is no putting it back. People are going to use AI chatbots. Because AI chatbots are getting good.

 

ChatGPT is hit and miss. When it hits it really saves a lot of time compared to a simple google search. Literally hours of work saved with one good prompt for some tasks, like finding out what framework/library to use and how to wrap it in your code.

When it misses, it's up to you to see chatgpt is bullshitting and refine the prompt or give up because that's something chatgpt can't do.

Bolded are critical points at this time. Bots are getting good. They aren't good yet. We are still several years out of where accuracy is in point that any bot could replace human interaction, even in simpler interactions (AFAIK bots lack nuance to steer conversation, they just react on what is given).

 

You need good prompt. Just like with google or even on forum. If your question doesn't work or your thread doesn't have enough information, getting answers will be difficult. One major advantage of forum is that people read and review post, then query for missing or partial information. Would bot have nuance to do that when prompt they are given lacks some critical information? I doubt it.

 

3 hours ago, 05032-Mendicant-Bias said:

It's a tool. There are things it can do well, and things it does badly. And often allucinates wrong answers. I wouldn't ask chatgpt to choose components for my new build, I would ask here in this forum. Chatgpt hasn't tried GPUs CPUs, etc... the users have. I would guess chatgpt training data is poisoned by all the fanboy/hater discussions about products to have a balanced understanding.

 

Luke on the wan show said he is now using chatgpt first instead of google when searching stuffs. You really are underestimating how good ChatGPT it is. And it's only going to get better from here on.

The data used is indeed quite big problem. Bot only has data its given. If we assume bot would have entire Google cache to work with, that still wouldn't be enough. It would have all opinions, tests, reviews and such, so yes, it would be able to give out very good answers based on those. If refined long enough, maybe even weight data to see what information is more accurate or opinion based than others.

 

But this would still only help those with fairly easy questions. Some data isn't on the internet. Some things might not get asked as often and not be cached because of that. Thats where human part, and experience comes in.

[Holmes108]

On 12/27/2022 at 5:50 PM, divito said:

The botphobia in this thread is concerning.

Right? I mean I agree in principle that I don't want it cluttering up threads. But the idea that we can distinguish the difference perfectly, and not confuse it with a persons style, or someone who is ESL or something is also concerning. 

 

Don't think we need to start dropping the ban hammer because a couple of people think something "seemed fishy". It's just a bit of a slippery slope. I do agree sometimes it can be kind of obvious, but I disagree that it's so cut and dry.

[Fendrick]

16 minutes ago, Holmes108 said:

Right? I mean I agree in principle that I don't want it cluttering up threads. But the idea that we can distinguish the difference perfectly, and not confuse it with a persons style, or someone who is ESL or something is also concerning. 

 

Don't think we need to start dropping the ban hammer because a couple of people think something "seemed fishy". It's just a bit of a slippery slope. I do agree sometimes it can be kind of obvious, but I disagree that it's so cut and dry.

Most are just phobic but also people like it here, in general humans at least the more emotional ones tend to hate anything that is not another human on the other end, I guess it may be because bots are not taught to reiterate 2000 times until they understand what was said perhaps?

 

Tribalism, a root cause for human success, hate anything that is not like us even if it could be what helps us.

Inefficiency over efficiency, just so people feel and look better.

 

People want us to be all multicultural, well you can see it is mostly a natural habit of the majority to hate what is not like them, survial mode.

 

https://www.mindcoolness.com/blog/tribalism-human-nature/

 

Now lets say we had a bot that no one detected as one, yet if you mention it's a bot!!! boom watch the emotions flow!

[Kid.Lazer]

6 hours ago, 05032-Mendicant-Bias said:

It's a tool. There are things it can do well, and things it does badly.

I think this is the defining characteristic, but not in a good way. It is indeed a tool, but it's so hit or miss that we don't even know if the tool will work in any given situation. That's like having a magic omni-use screwdriver that works in all types of screws... except for the ones it doesn't! And how do we know which one that is? It's a very wishy-washy, ethereal tool.

48 minutes ago, Fendrick said:

Most are just phobic

Can we please stop abusing the word "phobic?" (In this case specifically, but also in all areas of debate) Nobody is afraid of them, they just don't think their use is appropriate. I'm very tired of this mischaracterization.

[Fendrick]

Ok so their use is wrong, why is this discussion limited by naysayers to "Bots are bad at X and Y" But have no solutions? They also don't even bother observing human issues.

 

You can mark it down how you want, you shared your thoughts, but deep down people hate change and that is quite a change having to get your tech info from a bot.

 

Just think of how much free time people will have when they can no longer chat like this, it's actually far better than folks think it is.

[Beerzerker]

2 minutes ago, Kid.Lazer said:

I think this is the defining characteristic, but not in a good way. It is indeed a tool, but it's so hit or miss that we don't even know if the tool will work in any given situation. That's like having a magic omni-use screwdriver that works in all types of screws... except for the ones it doesn't! And how do we know which one that is? It's a very wishy-washy, ethereal tool.

Can we please stop abusing the word "phobic?" (In this case specifically, but also in all areas of debate) Nobody is afraid of them, they just don't think their use is appropriate. I'm very tired of this mischaracterization.

Agreed - If there is anything to worry about it's the possible intent behind it's useage. Such worries about it are not that it's a chatbot, it's the potential use it could be put to by the one behind it.

It's a tool and tools can be used for whatever purpose the user intends.
The "Whatever" about purpose being the sticking point.

As for me - I don't really care for it and to me, it's just unpersonal to use a bot like this in that you lose the human factor (Touch) that should be between you and the one using it.