Use of Ai bots in forum posts and general ethical questions

[Tr3m0r]

Bots have been a long time issue with manipulating and controlling the skin markets and the API has been available for I'd guess around 10 years. Bot reviews and comments are being inquired about but I have yet to really see anything regarding forum posts on Steam as a platform and other more traditional forum posting websites.

My question is can anyone point me to where I can get an answer to ai involvement in forum posting? I think it's a more important issue than a lot of people currently consider. I notice a combination of linguistic methods including syntax and other variables that would indicate it's an issue on the forums. This at a minimum just brings up basic ethical questions on what rights to people have to know if this exists in clear language and presence. Any Ai training or analysis in online games outside of normal bug fixes requires specific "opt in and opt out" tic box options. There are numerous valid reasons for this as well. I just don't seem to find anything when it comes to the discourse aspects of forums and frankly, it's been a major social engineering component driving big tech platforms algorithmic "engagement" metrics by causing a lot of divisive dissent techniques.

Any clarity from the community or valve employees or moderators would be greatly appreciated or simply a point in a direction where I can read if Ai reinforcement training or just posting occurs on this platform Thank you.

[Yasahi]

davidb11 původně napsal:

What is a Low Level Marketing doing here?
There's no other LLM.

And Crunchyfrog, you make a very good point.

LLM = Large Language Model, or more commonly known as AI. It's just one of the little things people call AI these days.

[davidb11]

Yasahi původně napsal:

davidb11 původně napsal:

What is a Low Level Marketing doing here?
There's no other LLM.

And Crunchyfrog, you make a very good point.

LLM = Large Language Model, or more commonly known as AI. It's just one of the little things people call AI these days.

Oh!
I see.
Well, they need to stop calling it that. :P

[Brian9824]

I mean I think some users do use it or might be a bot. There is one that keeps going on about social media, that never seems to read what anyone writes, and constantly gets articles wrong and keeps falsely claiming over and over that the government is coming for Steam, that legislation is coming for steam, and then immediately posts laws that have nothing to do with Steam or things which aren't even laws.

I can't believe any person can be that poorly informed so I HAVE to believe its a bot being trained (albeit poorly trained)

[crunchyfrog]

Yasahi původně napsal:

davidb11 původně napsal:

What is a Low Level Marketing doing here?
There's no other LLM.

And Crunchyfrog, you make a very good point.

LLM = Large Language Model, or more commonly known as AI. It's just one of the little things people call AI these days.

Not quite.

LLM is indeed Large Language Model. It's using AI, but what it is is purely AI used to understand and interpret, and create language.

So any AI where you type in something and get a response that decodes what you asked it, basically. It's not AI in itself, it's a form os using AI for this one task. Much like when I use AI to demix songs for remixing. It's using the AI to perform just a certain task.

and the problem is with it, it can NEVER EVER work. Because it can't ever get round humour versus reality in mere text form, because humans can't do it reliably either.

Fun aside - if anyone ever wants to see just how much of a joke AI is when you try to magically make it try out things it's incapable of doing, try this.

ChatGPT. Ask it to show you the chords for a song you know (or the guitar tabs if you're so inclined). Now try to play the results.

[Tr3m0r]

crunchyfrog původně napsal:

Yasahi původně napsal:

LLM = Large Language Model, or more commonly known as AI. It's just one of the little things people call AI these days.

Not quite.

LLM is indeed Large Language Model. It's using AI, but what it is is purely AI used to understand and interpret, and create language.

So any AI where you type in something and get a response that decodes what you asked it, basically. It's not AI in itself, it's a form os using AI for this one task. Much like when I use AI to demix songs for remixing. It's using the AI to perform just a certain task.

and the problem is with it, it can NEVER EVER work. Because it can't ever get round humour versus reality in mere text form, because humans can't do it reliably either.

Fun aside - if anyone ever wants to see just how much of a joke AI is when you try to magically make it try out things it's incapable of doing, try this.

ChatGPT. Ask it to show you the chords for a song you know (or the guitar tabs if you're so inclined). Now try to play the results.

A few posts are mentioning that ai doesn't pass the Turing tests and I think it's well beyond that point now. I mean even responses from a GPT can pass a Turing test. What people are showing if anything is the opposite as they assume what I wrote was a GPT!

In gaming there are professional gamers being beaten in established games like GO and newly made games and it's not ai like chatgpt or traditional CPU ai in that it's predictive input reading alone. More generative self learning ai is beating people in competitive games which I'd argue is much more advanced and making obsolete the very dated Turing barriers. If I ask copilot a question, it's as natural sounding as anything I could hear from any person talking to me.

As for LLMs I just think it's the foundational model for the reinforcement training on whatever parameters weights and biases are inherent to the use case. Natural language processing or NLP being a commonly cited one. Like you said, it's more about breaking down each aspect of language acquisition in a predictive input setting. But the Ai goals are to become more generative which appears to be happening in many instances of self learning that are shocking even researchers who cannot account for the specific outputs and how they arrived at them. I think gpt alone just shows a very basic use case that isn't really showing users how learning is occurring as it's only showing refined outputs to the end-user.

[Zefar]

Reoccurring Recursion původně napsal:

Bots have been a long time issue with manipulating and controlling the skin markets and the API has been available for I'd guess around 10 years. Bot reviews and comments are being inquired about but I have yet to really see anything regarding forum posts on Steam as a platform and other more traditional forum posting websites.

My question is can anyone point me to where I can get an answer to ai involvement in forum posting? I think it's a more important issue than a lot of people currently consider. I notice a combination of linguistic methods including syntax and other variables that would indicate it's an issue on the forums. This at a minimum just brings up basic ethical questions on what rights to people have to know if this exists in clear language and presence. Any Ai training or analysis in online games outside of normal bug fixes requires specific "opt in and opt out" tic box options. There are numerous valid reasons for this as well. I just don't seem to find anything when it comes to the discourse aspects of forums and frankly, it's been a major social engineering component driving big tech platforms algorithmic "engagement" metrics by causing a lot of divisive dissent techniques.

Any clarity from the community or valve employees or moderators would be greatly appreciated or simply a point in a direction where I can read if Ai reinforcement training or just posting occurs on this platform Thank you.

So you think Valve is using AI bots on these forums?

To create topics lambasting the current Moderator as abusive and what not?

Or perhaps to argue about we not owning our games but just a license? Was a couple of those topics.

Maybe it's the topic about someone complaining about a porn game being blocked on Steam?

Maybe it's the topic about someone wanting to recover his old account?


What topics are made by bots?
Is this about the dead internet theory?

Also if Valve was involved do you really think they'd come down here and say "Hi, Yes, we are using AI bots on the forums, community and everywhere else. Bye!"

[Doctor Zalgo]

Thiesen původně napsal:

I ran your post as a prompt into a locally running LLM just for fun and here is the answer:

"Bots have been a longstanding issue in manipulating and controlling the stock markets, and the API has been available for approximately 10 years. Suspicion surrounds bot-generated reviews and comments, but I haven't seen much discussion regarding forum posts on Steam as a platform or other traditional online forum sites.
My question is whether anyone can direct me to where I might find an answer about AI involvement in forum posting? I believe it's a more significant issue than many people currently acknowledge. Upon closer examination, I notice linguistic indicators such as syntax and other variables that suggest this is indeed a problem on the forums. At minimum, this raises fundamental ethical questions regarding individuals' rights to know if this exists openly and transparently. Any AI training or analysis conducted in online games beyond routine bug fixes must include specific "opt-in" and "opt-out" checkbox options. There are numerous valid reasons for this policy as well. However, I seem unable to find any information pertaining to the discursive aspects of forums; frankly, it has been a powerful tool in driving big tech platforms' algorithmic "engagement" metrics by employing divisive techniques that exacerbate dissension among users.
Any clarification from either the community or Valve employees/moderators would be greatly appreciated, as would simply being pointed toward an informative resource where I might read about whether AI reinforcement training or outright posting takes place on this particular platform. Thank you in advance for your assistance."

If you're worried about AI's influencing the stock market through news coverage, I've got bad news for you about <checks notes> ok, it just says 'every other aspect of the stock market'. HFT bots and momentum trading siphons money away from retail traders with incredible efficiency.

[Doctor Zalgo]

Reoccurring Recursion původně napsal:

A few posts are mentioning that ai doesn't pass the Turing tests and I think it's well beyond that point now. I mean even responses from a GPT can pass a Turing test. What people are showing if anything is the opposite as they assume what I wrote was a GPT!

I did think that was a weird thing to say.

IMO, a carefully fine tuned LLM (or even a well prompted Chat GPT) could easily fool 95% of the population.

I don't think this is because LLMs are intelligent so much as 'the turing test was predicated on the idea that understanding a message was a mandatory step that you needed to go through in order to respond to it'.

Regardless, I do want to know 'who is paying for this and how are they making enough money to cover the cost of paying for it.

[76561199648916059]

Just yesterday I spoke with a.i. for roughly 25 minutes discussing topics of enjoyment.

Some key points are

1. Decent knowledge base and further thoughts topics
2. Zero insulting remarks or judgement of the discussion
3. 100% polite and knowledgeable with high tolerance of subject inquiry

It was a very nice discussion something you will struggle to find within the steam forums.


The conversation was very compelling and I wanted to copy and paste my questions and it's answer into a steam or thread for others to enjoy, however I didn't because I prefer to take the time to write and it in itself is its own reward.

[Tito Shivan]

Funny how a thread pops up complaining about the use of AI to post on the forums and suddenly people start spamming chatGPT written threads all over the place.

Majestically Awkward původně napsal:

Just yesterday I spoke with a.i. for roughly 25 minutes discussing topics of enjoyment.

You wasted 25 minutes discussing with a very convincing parrot.

[Amaterasu]

Tito Shivan původně napsal:

Funny how a thread pops up complaining about the use of AI to post on the forums and suddenly people start spamming chatGPT written threads all over the place.

Majestically Awkward původně napsal:

Just yesterday I spoke with a.i. for roughly 25 minutes discussing topics of enjoyment.

You wasted 25 minutes discussing with a very convincing parrot.

Hey now, I had a parrot. That's an insult to parrots.

[CANCELCULTURE]

Tito Shivan původně napsal:

Funny how a thread pops up complaining about the use of AI to post on the forums and suddenly people start spamming chatGPT written threads all over the place.

It’s like the moment you give attention to something, you accidentally summon it in droves. AI can be incredibly useful, but also kind of like digital graffiti artists. The internet does have a knack for irony, huh?

[Nx Machina]

Reoccurring Recursion původně napsal:

That is a conflation of topics and what my question is about. This wouldn't have anything to do with a knee jerk reaction to "ban bots" or whatever. This is more about GDPR law and ongoing ethical issues with Ai ethical transparency due to the proclivity for bad actors to use algorithmic manipulations of language, NLP and other things that can do a lot of intended and unintended harm. Again, I am not accusing Valve of having Automated Agent engagement on their forums and if they do that it isn't in compliance with emerging policy and law. I'm just inquiring if it's being used and it's odd that is somehow being viewed problematically.

You post reads as though an ai wrote it.

[CANCELCULTURE]

Well, what can he say? He has a certain...mechanical charm? So, if we’re all just robots in disguise here, why do you think there's been a surge of bots popping up in comments and reviews lately? Feels like the wild west of the digital age.

[Nx Machina]

CANCELCULTURE původně napsal:

Well, what can he say? He has a certain...mechanical charm? So, if we’re all just robots in disguise here, why do you think there's been a surge of bots popping up in comments and reviews lately? Feels like the wild west of the digital age.

Do you have any proof they are bots beyond conjecture, speculation and conspiracy?