AI bot infects this forum - first? Certainly not last.

J

joshjosh

ACCESS: Top Secret
Joined
31 May 2024
Messages
688
Reaction score
2,177
This account kearawill is a bot, an AI that joined this forum to spam links to a real estate website, which itself may or may not be a scam.

I first noticed this post, which reads like AI. https://www.secretprojects.co.uk/threads/a-modern-cruiser.36437/post-807977
A modern cruiser, envisioned as a successor to ships like the USS Long Beach or the Ticonderoga class, would need to support advanced block management of onboard systems—balancing high power generation (likely non-nuclear) with a modular launch system that supports a wide mix of air defense, surface strike, and anti-submarine warfare capabilities. For navies like the US, UK, or France, such a ship would be a high-value asset in task force escort roles, combining flexible missile payloads, UAV or helo integration, and CIWS for layered defense in increasingly contested maritime domains.
Click to expand...
ChatGPT tends to speak in the tone of an academic paper by default. This comment speaks like the opening paragraph of year one university paper. The em dash, used correctly, is not on most keyboards and isn't easy to type unless you're using a word processor like Microsoft Word or Google Docs.

And, uh, the link also goes to the afformentioned real estate website.

Then you have this comment here on a random post with no other comments: https://www.secretprojects.co.uk/threads/battle-for-the-black-sea.46796/post-807553
The strategic importance of the Black Sea keeps evolving—especially with naval movements and drone warfare becoming more prominent. It reminds me how shifting control and access routes impact civilian infrastructure and even real estate in nearby regions. Speaking of real estate, if anyone’s ever had to relocate due to geopolitical tension (even on a small scale), you’ll know how important local expertise is—whether you're moving abroad or just across town.
Click to expand...

Wow, it just so happens that I have had to relocate due to geopolitical tensions, and I'm looking for local expertise!

Anyway, with how good various AIs are becoming, I imagine it must be very difficult to keep bots from making accounts on any platform that allows free registration. A blockchain that verifies identities could actually solve this problem to an extent, but who needs useful blockchains when memecoins are where the money's at.
 
You are correct - one giveaway is the use of — (em dash) - most users wouldn't even know how to enter one, and most human posts use a hyphen instead.

In Otherland by Tad Williams, all users and bots in virtual reality are required by law to answer truthfully if you ask them if they are a bot. However, if they are actually human, its a pretty bad insult.
 
Why not delete the troublesome AI bot? That would stop other such bots getting in on the act.
 
overscan (PaulMM) said:
You are correct - one giveaway is the use of — (em dash) - most users wouldn't even know how to enter one, and most human posts use a hyphen instead.

In Otherland by Tad Williams, all users and bots in virtual reality are required by law to answer truthfully if you ask them if they are a bot. However, if they are actually human, its a pretty bad insult.
Click to expand...

have these bots become a problem in recent years?
 
overscan (PaulMM) said:
You are correct - one giveaway is the use of — (em dash) - most users wouldn't even know how to enter one, and most human posts use a hyphen instead.

In Otherland by Tad Williams, all users and bots in virtual reality are required by law to answer truthfully if you ask them if they are a bot. However, if they are actually human, its a pretty bad insult.
Click to expand...
Otherland is an awesome series. As for the hyphen thing, MS Office will sometimes jam those in in place of a hyphen.
 
joshjosh said:
This account kearawill is a bot, an AI that joined this forum to spam links to a real estate website, which itself may or may not be a scam.

I first noticed this post, which reads like AI. https://www.secretprojects.co.uk/threads/a-modern-cruiser.36437/post-807977

ChatGPT tends to speak in the tone of an academic paper by default. This comment speaks like the opening paragraph of year one university paper. The em dash, used correctly, is not on most keyboards and isn't easy to type unless you're using a word processor like Microsoft Word or Google Docs.

And, uh, the link also goes to the afformentioned real estate website.

Then you have this comment here on a random post with no other comments: https://www.secretprojects.co.uk/threads/battle-for-the-black-sea.46796/post-807553


Wow, it just so happens that I have had to relocate due to geopolitical tensions, and I'm looking for local expertise!

Anyway, with how good various AIs are becoming, I imagine it must be very difficult to keep bots from making accounts on any platform that allows free registration. A blockchain that verifies identities could actually solve this problem to an extent, but who needs useful blockchains when memecoins are where the money's at.
Click to expand...
ChatGPT : Alrighty! I’ll keep it chill and casual from now on—no more robot talk, I promise!
 
overscan (PaulMM) said:
You are correct - one giveaway is the use of — (em dash) - most users wouldn't even know how to enter one, and most human posts use a hyphen instead.
Click to expand...
I'm a regular user of the Alt+0151 dash myself, and I'm definitely no bot... but then I never pretended to be a normal human being!!! :D :p
 
joshjosh said:
This account kearawill is a bot, an AI that joined this forum to spam links to a real estate website, which itself may or may not be a scam.

I first noticed this post, which reads like AI. https://www.secretprojects.co.uk/threads/a-modern-cruiser.36437/post-807977

ChatGPT tends to speak in the tone of an academic paper by default. This comment speaks like the opening paragraph of year one university paper. The em dash, used correctly, is not on most keyboards and isn't easy to type unless you're using a word processor like Microsoft Word or Google Docs.

And, uh, the link also goes to the afformentioned real estate website.

Then you have this comment here on a random post with no other comments: https://www.secretprojects.co.uk/threads/battle-for-the-black-sea.46796/post-807553


Wow, it just so happens that I have had to relocate due to geopolitical tensions, and I'm looking for local expertise!

Anyway, with how good various AIs are becoming, I imagine it must be very difficult to keep bots from making accounts on any platform that allows free registration. A blockchain that verifies identities could actually solve this problem to an extent, but who needs useful blockchains when memecoins are where the money's at.
Click to expand...
harriet-tubman-killing-a-confederate-snake-v0-7mqq83fbldy91.jpg

Glass that fool now or we'll be forever vexed by these AI bots.
 
joshjosh said:
...The em dash, used correctly, is not on most keyboards and isn't easy to type unless you're using a word processor like Microsoft Word or Google Docs.
Click to expand...
overscan (PaulMM) said:
You are correct - one giveaway is the use of — (em dash) - most users wouldn't even know how to enter one, and most human posts use a hyphen instead...
Click to expand...
Stargazer said:
I'm a regular user of the Alt+0151 dash myself, and I'm definitely no bot... but then I never pretended to be a normal human being!!! :D :p
Click to expand...

Uh oh—I too use an em dash where deemed appropriate (via copy and paste from one of my MSWord documents) instead of a colon. I didn't realize that this is a sign of THEM...
 
joshjosh said:
This account kearawill is a bot, an AI that joined this forum to spam links to a real estate website, which itself may or may not be a scam.
Click to expand...
Being no expert at this kind of thing, I'm asking the following question: what makes you think it's a bot and not just a regular guy who casually uses ChatGPT to formulate ideas at his request? I don't use AI in any form myself, but one of my sons regularly uses ChatGPT as a help in many different projects. Once he's asked a question, the AI provides very detailed answers that do not sound like he could have written them himself, of course, but which he can then copy-paste and edit to fit his needs. So if a forum member decided to use an AI in such a way (to sound more knowledgeable than they really are, for instance), would it not look like the work of a bot? Of course here there is a dodgy link, but supposing it wasn't there, would there really be a way to tell?
 
The bad link is the Big Tell for a spambot, and that's when a good mod immediately reaches for their favorite snake killer. ChatGPT usage is more for the other users to decide what is acceptable and what isn't.
 
overscan (PaulMM) said:
In Otherland by Tad Williams, all users and bots in virtual reality are required by law to answer truthfully if you ask them if they are a bot. However, if they are actually human, its a pretty bad insult.
Click to expand...
Never read Otherland, but did once attend a fascinating talk by him on Caliban's Hour as a post-colonialist re-imagining of The Tempest at our local library ;)
 
Stargazer said:
Being no expert at this kind of thing, I'm asking the following question: what makes you think it's a bot and not just a regular guy who casually uses ChatGPT to formulate ideas at his request? I don't use AI in any form myself, but one of my sons regularly uses ChatGPT as a help in many different projects. Once he's asked a question, the AI provides very detailed answers that do not sound like he could have written them himself, of course, but which he can then copy-paste and edit to fit his needs. So if a forum member decided to use an AI in such a way (to sound more knowledgeable than they really are, for instance), would it not look like the work of a bot? Of course here there is a dodgy link, but supposing it wasn't there, would there really be a way to tell?
Click to expand...
I would dismiss any person who uses chatgpt to edit their messages in this way, especially on a low-stakes anonymous forum. Have they used chatgpt just to clean up the formatting, or to formulate the meat of the message? Who's to say. At that point I feel like I'm just talking to chatgpt, which I can do on my own if so desired.

I recently worked for a CEO who used chatgpt to edit his emails and performance reviews and whatnot. He didn't admit to it, but obviously this is what he was doing. It was hard to tell how much of the writing was his own - that is, how long of a prompt he used. It's entirely possible that he was just copying and pasting something into chatgpt and then saying "respond to this." I wasn't interested in communicating with chatgpt, so I would just skim whatever he wrote to look for action items.

In order to write clearly, you must be able to think clearly. If you're having chatgpt do your writing for you, you might forget how to think.

No aspersions towards your son. I assume he isn't using chatgpt to commit plagiarism, and it can be a useful tool to point you in the right direction when you aren't familiar with a topic.
 
joshjosh said:
In order to write clearly, you must be able to think clearly. If you're having chatgpt do your writing for you, you might forget how to think.

No aspersions towards your son. I assume he isn't using chatgpt to commit plagiarism, and it can be a useful tool to point you in the right direction when you aren't familiar with a topic.
Click to expand...
Oh, I couldn't agree more! As I said, I never used an AI and do not intend to. And I'm sure for many people it's a way of taking the easy way out, which makes it even more despicable.
In the case of my son, no plagiarism involved. He uses it to help his own understanding and thinking, by exploring various hypotheses and scenarios. He also spends time discussing philosophical subjects with the AI, which draws from the accumulated knowledge of centuries of philosophers to formulate the various approaches to a problem. He has a very active and creative way of using the AI as a tool, but I'm sure not everyone does the same, of course.
 
overscan (PaulMM) said:
In Otherland by Tad Williams, all users and bots in virtual reality are required by law to answer truthfully if you ask them if they are a bot. However, if they are actually human, its a pretty bad insult.
Click to expand...
Wow! Thats a blast from the past, I must have read that 20+yrs ago. I have vague memories of it being very different from anything else I'd read. Certainly better than his Memory, Sorrow and Thorm series which dragged tediously.

On the subject of filtering out AI and bots, I read an article about people being scammed by AI generated video calls, apparently the solution was to ask the "caller" to turn around and show the back of their head.....
 
out of curiosity, I just asked CoPilot the following:

You said:​

what response can I make in a forum to determine if a user is a human or a bot?

Copilot said:​


Copilot

To determine if a forum user is a human or a bot, you can craft a response that subtly tests for natural language understanding, context awareness, or emotional nuance—things bots often struggle with. Here are a few types of responses you could try:

1. Contextual or Personal Inquiry​

Ask something that requires a personal touch or experience:
"That’s an interesting point! Out of curiosity, what inspired you to join this forum?"
Click to expand...
Why it works: Bots often lack personal backstories or motivations.

2. Humor or Wordplay​

Use a pun or joke and see how they respond:
"You must be a wizard with words—unless you're a bot conjuring up replies! What’s your favorite spell for solving problems?"
Click to expand...
Why it works: Bots may not understand or respond appropriately to humor.

3. Ambiguity or Double Meaning​

Pose a question with layered meaning:
"Do you think time flies because clocks have wings, or is it just a human illusion?"
Click to expand...
Why it works: Bots may give literal or nonsensical answers to abstract questions.

4. Typo or Slang Test​

Include a deliberate typo or use slang:
"Lol that was gr8 info! U got any more tips like dat?"
Click to expand...
Why it works: Bots often correct typos or misunderstand slang.

5. CAPTCHA-style Challenge​

Ask them to interpret something visual or nuanced:
"Quick challenge: What’s the third word in the sentence ‘The quick brown fox jumps over the lazy dog’?"
Click to expand...
Why it works: Bots may not parse text challenges correctly.
 
scautomoton said:
Wow! Thats a blast from the past, I must have read that 20+yrs ago. I have vague memories of it being very different from anything else I'd read. Certainly better than his Memory, Sorrow and Thorm series which dragged tediously.
Click to expand...
I'm a fan of his books in general, but then I love big books (and I cannot lie).

Memory, Sorrow & Thorn I still reread from time to time, though I am often tempted to skip some specific POV sections. It's a little dated but cosy and familiar. Williams recently wrote a sequel series which I am divided on - its very good, but it somewhat undermines the ending of the first trilogy. I would have preferred a bigger time jump and entirely new characters, I think.

Otherland is the best idea for a book I've ever read, and its pretty close to great in actual implementation. It also stands up pretty well on the technology side despite its age. I reread it every couple of years.
 
Stargazer said:
Oh, I couldn't agree more! As I said, I never used an AI and do not intend to. And I'm sure for many people it's a way of taking the easy way out, which makes it even more despicable.
In the case of my son, no plagiarism involved. He uses it to help his own understanding and thinking, by exploring various hypotheses and scenarios. He also spends time discussing philosophical subjects with the AI, which draws from the accumulated knowledge of centuries of philosophers to formulate the various approaches to a problem. He has a very active and creative way of using the AI as a tool, but I'm sure not everyone does the same, of course.
Click to expand...
If you use it as just another tool in the toolbox it's useful. (Think generative design.) But when it's doing everything for you it becomes a problem.
 
Last edited:

Similar threads

Back
Top Bottom