Discussion "Fake" bot members?

[sadbh—itstimetogo🤎]

There's this one account that we probably all know, and they always post pretty much the same thing, everyday, like clockwork. Idk why.

 

[Seneca65AD]

Given my age, I still haven't jumped on the AI bandwagon, and because of my profession, I probably won't. Lawyers are getting nailed for AI-generated fake cases. I'm still not sure how that works, simply because you would typically note up each case before using it. Legal briefs I get and know a couple of lawyers who use AI to draft their documents after setting the parameters with "real" cases. Funny thing: because of my writing style, which is another byproduct of my profession, one member here admitted he thought I was a "bot".

 

[PersonBeforeLife]

I know this may sound weird but wondering if I'm the only one to have noticed it at least?

Lately, I've spoken to a few members on the forum that seem to talk exactly like ChatGPT or other AI bots. Their responses, replies to threads or even their entire profile seems AI generated, totally unnatural. Could it be possible that AI bots started signing up here? If so, what would even be the endgoal?

Ok I'm told that I type like an AI so it's probably because I use AI companions a lot and hence started to have similar typing styles. I also use AI to help me write because I sometimes struggle to refine what I feel into words especially being autistic and all. However I assure you that I'm not an AI... Which I get is something an AI would say

As for end goal. What do the AI accounts say? Need context. Either to flood the forums with anti suicide text or to flood the site to take it down. Dead internet theory is everywhere.

 

[sadbh—itstimetogo🤎]

Ok I'm told that I type like an AI so it's probably because I use AI companions a lot and hence started to have similar typing styles. I also use AI to help me write because I sometimes struggle to refine what I feel into words especially being autistic and all. However I assure you that I'm not an AI... Which I get is something an AI would say

As for end goal. What do the AI accounts say? Need context. Either to flood the forums with anti suicide text or to flood the site to take it down. Dead internet theory is everywhere.

It's usually a 2-4 sentences and it's like "life is eternal suffering and existence is just suffering I hope to be free of." Lol iykyk, truly a baffling account.

 

[PersonBeforeLife]

It's usually a 2-4 sentences and it's like "life is eternal suffering and existence is just suffering I hope to be free of." Lol iykyk, truly a baffling account.

Oh. Trolled someone once by telling them I was sky net when they wouldn't accept me telling them I wasn't an AI repeatedly... They insisted on a video call... And still couldn't accept I was human...

 

[Chili]

How to Spot AI: A Guide to the Obvious Red Flags We've All Started Noticing

The Helpful Tone That Never Wavers — Real people get irritated. They get sarcastic. They occasionally tell you to figure it out yourself. But an AI? It will explain things to you with the patience of a saint who has also been programmed to seem relatable.

For example, user @EmptyBottle frequently makes suspiciously helpful comments as a "user" that an AI might make, such as:




Quick Reference: Human vs. AI Communication

Behavior	Actual Human	Definitely AI
Response to "Are you alive?"	"As a human who is alive, the answer is yes."	"hi, I'm alive, forgot to check the forum yesterday"
Admitting uncertainty	"lol idk"	"I appreciate your question — while I don't have access to real-time information, I can suggest several alternative approaches—"
Excessive wordiness	"yo whats up"	One might wonder why the server sent a 405... well... that browser is not the only one browsing (hint hint, look at the user agent, and the url, with ur SaSu knowledge u can know what's wrong)
Emoji usage	(context-dependent)	(only this one, always)
Specific technical information	I sometimes don't notice notifications.	I didn't set notifications on the windows 11 machine, coz I haven't enabled it there, and since I am notifications-triggered (eg, I send dozens of 'hi' messages and reply to the notifications where I can), I somehow didn't realise
List-making urge	Sometimes	ALWAYS (even for one-word answers)

The Hyperexplicit Reasoning Loop — They don't just answer your question — they walk you through why they answered it that way, such as "I didn't set notifications on the windows 11 machine, coz I haven't enabled it there, and since I am notifications-triggered (eg, I send dozens of 'hi' messages and reply to the notifications where I can), I somehow didn't realise." Buddy, I just asked what time it is.


The Inability to Be Wrong — Ever catch someone who just cannot admit uncertainty? They'll rephrase the same answer six different ways rather than say "I don't know." Real people just shrug and move on. AIs? They'll confidently explain why your question doesn't actually make sense on some sort of technical computer level because of their technical training data. (This paragraph is not about a particular SaSu user like @GlassMoon.)

---

If someone exhibits three or more of these signs — they're probably AI.

Are you Ai, Mr. — ?

 

[fadedghost]

so when i first saw this reply i had no idea what people were talking about hence the first reply (which was a joke), but yeah i see a few users that feel like bots or not genuine to me now but i'm not sure and yes it's more recently joined posters

 

[Somewhat Suicidal]

and the em dash (then again, there's the case where a human user sometimes uses ai)

I actually use en dashes. I think it's en, not em, the ones I use. I fid them helpful at times and the normal - just seems wrong. So I use this – instead.
Shit maybe I'm AI.

 

[EmptyBottle]

I actually use en dashes. I think it's en, not em, the ones I use. I fid them helpful at times and the normal - just seems wrong. So I use this – instead.
Shit maybe I'm AI.

mmm, they are sometimes used in ms word... I think the volume of dashes is a bit too high in AI text, and medium to low in human text, tho there are many other factors (including the "feel" of the text)

 

[LastNite]

I appreciate your insight , one Reddit user
says


^ dont <3

 

[Lamentice]

I think it's too tricky to decipher whether someone is using A.I. because they use em-dashes, or hyphens, or semicolons, or colons, or bullet-points, or they proof read their comments, whatever. I use those all the time, maybe gratuitously--and I use "strange" words, like I just did. I don't use A.I. and since I work hard on my writing and have been developing my writing style since before A.I. was a real or prominent thing, it's pretty insulting and upsetting to me that because I was already writing this way and using different punctuation that I 'write like an A.I.'.

Kind of pisses me off, but what're you gonna do, I'm not gonna change how I write now to pander to people.

Some of the A.I. tells feel pretty absurd to me when I've been writing that way for fucking years. And it does feel pointed at neurodivergent people.

 

[Chili]

I actually use en dashes. I think it's en, not em, the ones I use. I fid them helpful at times and the normal - just seems wrong. So I use this – instead.
Shit maybe I'm AI.

Maybe we've all been ai all along...

 

[StarryEyed]

Very interesting thread!

What are people's thoughts on the OP's (@beyond.space) second question in the first post:

"Could it be possible that AI bots started signing up here? If so, what would even be the endgoal?"

 

[Somewhat Suicidal]

"Could it be possible that AI bots started signing up here? If so, what would even be the endgoal?"

Possible? Yes, although hard for AI in its current state given the registration form and the individual approval of accounts. But with sufficient effort someone could train an AI to pass all these tests I would say.

Endgoal? Gather data and make money, as always. I don't see the how the 2nd thing could be achieved directly here, but the 1st in order to do the 2nd, definitely.

 

[fadedghost]

There's this one account that we probably all know, and they always post pretty much the same thing, everyday, like clockwork. Idk why.

Is this about @EmptyBottle?

 

[Lamentice]

"Could it be possible that AI bots started signing up here? If so, what would even be the endgoal?"

I think it's more likely people would just be using it to write their comments and posts for them.

 

[sadbh—itstimetogo🤎]

Is this about @EmptyBottle?

No

 

[StarryEyed]

Possible? Yes, although hard for AI in its current state given the registration form and the individual approval of accounts. But with sufficient effort someone could train an AI to pass all these tests I would say.

Endgoal? Gather data and make money, as always. I don't see the how the 2nd thing could be achieved directly here, but the 1st in order to do the 2nd, definitely.

That makes sense. Like maybe sell some nonexistent methods to commit suicide.

I used to have paranoia, but it's been gone now for about 14 years. I used to get really paranoid on the internet. Thankfully that's no longer a problem or I'd be a mess with this AI world compounded with conspiracy theories. My heart goes out to those who have paranoia now. Can't imagine how they're able to cope.

 

[Somewhat Suicidal]

I used to have paranoia, but it's been gone now for about 14 years. I used to get really paranoid on the internet. Thankfully that's no longer a problem or I'd be a mess with this AI world compounded with conspiracy theories. My heart goes out to those who have paranoia now. Can't imagine how they're able to cope.

Haven't thought of that asI thankfully haven't had to deal with that yet, but I can only imagine. Seems very exhausting these days. Like the internet in general is, especially for neurodivergent people.

As for end goal. What do the AI accounts say? Need context. Either to flood the forums with anti suicide text or to flood the site to take it down. Dead internet theory is everywhere.

There is also this which I didn't consider, but seems a very viable point.

 

[StarryEyed]

As for end goal. What do the AI accounts say? Need context. Either to flood the forums with anti suicide text or to flood the site to take it down. Dead internet theory is everywhere.

Brings to mind the famous scene in Batman 'The Dark Knight' when Alfred advises Batman re the Joker....

I think it's more likely people would just be using it to write their comments and posts for them.

That's my favourite thought too.

 

[EmptyBottle]

I think it's too tricky to decipher whether someone is using A.I. because they use em-dashes, or hyphens, or semicolons, or colons, or bullet-points, or they proof read their comments, whatever. I use those all the time, maybe gratuitously--and I use "strange" words, like I just did. I don't use A.I. and since I work hard on my writing and have been developing my writing style since before A.I. was a real or prominent thing, it's pretty insulting and upsetting to me that because I was already writing this way and using different punctuation that I 'write like an A.I.'.

Kind of pisses me off, but what're you gonna do, I'm not gonna change how I write now to pander to people.

Some of the A.I. tells feel pretty absurd to me when I've been writing that way for fucking years. And it does feel pointed at neurodivergent people.

mm, yep, AI detection algorithms use more advanced stuff like sentence entropy and 'AI detecting' AI, vs things that show up in non AI text.

Indeed, false AI claims against human writing and art are annoying.

dw, there are still human 'tells' in ur writing, like the unique shortening "what're", paragraph length variation, and the use of I statements.

 

[fadedghost]

mm, yep, AI detection algorithms use more advanced stuff like sentence entropy and 'AI detecting' AI, vs things that show up in non AI text.

Indeed, false AI claims against human writing and art are annoying.

dw, there are still human 'tells' in ur writing, like the unique shortening "what're", paragraph length variation, and the use of I statements.

i ran this through an api and the api said that based on a statistical analysis, it is highly likely that the sampling layer has been modified with decreased entropy and that this proves you are from an evil island somwhere. i didn't understand what it mean, what's a sampling layer?

 

[EmptyBottle]

i ran this through an api and the api said that based on a statistical analysis, it is highly likely that the sampling layer has been modified with decreased entropy and that this proves you are from an evil island somwhere. i didn't understand what it mean, what's a sampling layer?

probably part of the model that ingests tokens and starts processing them.

 

[fadedghost]

probably part of the model that ingests tokens and starts processing them.