AI Bot Accounts - things to look for.

[midicon]

I'm not really fussed lol. I doubt anyone could ever imagine me being a Bot.

There is one poster though, that I sometimes think is one.

We might all have different people in mind!

[Kurokage]

I'm not really fussed lol. I doubt anyone could ever imagine me being a Bot.

There is one poster though, that I sometimes think is one.

We might all have a different user in mind!

I think all members should pinch themselves right now to make sure they aren't the robots!!

  ​​​​​​​

​​​​​​​Beep boop beep......

[ReturnofBroccoli]

I'm not really fussed lol. I doubt anyone could ever imagine me being a Bot.

There is one poster though, that I sometimes think is one.

Is it me were you there last night when I got asked? Its like the 6th user that has asked me and I cant figure out what im doing thats making them think this way

[midicon]

Is it me were you there last night when I got asked? Its like the 6th user that has asked me and I cant figure out what im doing thats making them think this way

Not you, you are too cool lol!

[ReturnofBroccoli]

Not you, you are too cool lol!

  Broccoli's head grew 3 sizes today
I would have said my Florets but it sounds perverted

Make room cool kids table broccoli strut

[Halfswede]

I'm curious to know the account you speak of but T&C and all.
Is it possible that this user is just using some LLM for all their responses? This is very common and will only become even more so. That doesn't make them a bot, just someone using a tool. Though if they are using it for every response they are doing themselves no favors.

Using an LLM for responses is basically being a hybrid bot. You are asking the thing to create your response by feeding it the initial post. If you didn't feed it the original post, then that is different.

There is the only real exception and that is if you said 'spell and grammar check this', but if someone is using it to generate the response, that's literally a bot. As soon as you add anything else, that's not you even if you think you are driving. Just be a human and participate.

Whether or not there are bot accounts isn't even a question. There are even flow-diagrams that have been intercepted showing exactly how they are designed to respond to certain situations. They are on every message board and forum in existence except maybe knitting. If a site  has anything controversial going on, it is a direct target for these things. The algorithms have been refined to such a degree that bots are successfully passing job interviews, getting hired for WFH jobs and successfully getting paid (for somebody).  

What people forget is that they only really got refined in the last couple of years. 20218 was the first public LLM let loose and they only got good in the last year or so. People had to be involved.

The thing that triggered my curiosity about a particular user was a confusingly nonsensical post following up to a pretty clear comment a user posed. It was like if you didn't read in depth, it fit, but was actually just complete, kamala harris level, word-salad, fluffery. So I started looking at all of the posts and there were several distinct patterns.

They are 100% there, it's just sad that we get duped into conversing with them.  To be fair though, most of the people here, even the ones that are accused of shilling (big difference from AI bot), are actual humans.  Some people literally get paid or are given time-off a sentence for shilling propaganda.

[Halfswede]

To be clear, I have seen many likely paid or incentivized shills come and go on ATS and here, but this was the only time I think I found and actual non-human account that was trying to pass as human and not some spammy garbage bot.

[ReturnofBroccoli]

Using an LLM for responses is basically being a hybrid bot. You are asking the thing to create your response by feeding it the initial post. If you didn't feed it the original post, then that is different.

There is the only real exception and that is if you said 'spell and grammar check this', but if someone is using it to generate the response, that's literally a bot. As soon as you add anything else, that's not you even if you think you are driving. Just be a human and participate.

Whether or not there are bot accounts isn't even a question. There are even flow-diagrams that have been intercepted showing exactly how they are designed to respond to certain situations. They are on every message board and forum in existence except maybe knitting. If a site  has anything controversial going on, it is a direct target for these things. The algorithms have been refined to such a degree that bots are successfully passing job interviews, getting hired for WFH jobs and successfully getting paid (for somebody).  

What people forget is that they only really got refined in the last couple of years. 20218 was the first public LLM let loose and they only got good in the last year or so. People had to be involved.

The thing that triggered my curiosity about a particular user was a confusingly nonsensical post following up to a pretty clear comment a user posed. It was like if you didn't read in depth, it fit, but was actually just complete, kamala harris level, word-salad, fluffery. So I started looking at all of the posts and there were several distinct patterns.

They are 100% there, it's just sad that we get duped into conversing with them.  To be fair though, most of the people here, even the ones that are accused of shilling (big difference from AI bot), are actual humans.  Some people literally get paid or are given time-off a sentence for shilling propaganda.

People get paid to shill? ..... lol..... where does one sign up for this job I could use the money. Shills'R'us.com I think I could shill if I tried. Having someone pick the side I debate on I cant even choose what im having for dinner. Someone ask chatgpt7000 if shilling is profitable maybe we can form some kind of shill firm hahaha

All joking aside does ToS even apply concerning bots 0.o

[Halfswede]

  Broccoli's head grew 3 sizes today
I would have said my Florets but it sounds perverted

Make room cool kids table broccoli strut

This is a silly question, but more for my curiosity. First, it's clear you are no bot and not just for the many reasons in your post. 

Just posting in this thread though makes it fun to analyze people's posts. When you were typing out 'florets', did your autocorrect put 'Florence' and you had to edit it?  This is more one of those questions of did I Sherlock Holmes this one, or was it just accidentally capitalized?

[ReturnofBroccoli]

This is a silly question, but more for my curiosity. First, it's clear you are no bot and not just for the many reasons in your post. 

Just posting in this thread though makes it fun to analyze people's posts. When you were typing out 'florets', did your autocorrect put 'Florence' and you had to edit it?  This is more one of those questions of did I Sherlock Holmes this one, or was it just accidentally capitalized?

No I didnt, I think i may have capitalized it myself actually Florets nope just tested it my mobile autocorrected to it

Capitalized it automatically I mean