What is a bot and how do they infiltrate forums like this

[Redmires]

I've seen a few posts on here now that are reckoned to be generated by 'bots'. How does a bot operate and how do they generate content. A user has to sign up and confirm membership via an email address etc. I'm not having a pop at this forum but there must be some kind of anti-bot technology to stop them from infiltrating excellent forums such as TLF. Perhaps the signing up process needs to be more stringent - 2FA maybe. Or do the bots have mobile phones as well

[Mike4]

I've seen a few posts on here now that are reckoned to be generated by 'bots'. How does a bot operate and how do they generate content. A user has to sign up and confirm membership via an email address etc. I'm not having a pop at this forum but there must be some kind of anti-bot technology to stop them from infiltrating excellent forums such as TLF. Perhaps the signing up process needs to be more stringent - 2FA maybe. Or do the bots have mobile phones as well

On another forum I use, there was a bit of a kerfuffle over regular users developing a habit of calling out new posters making generic bot-like posts, as bots. Team Mod asked us to stop it as they were vetting all new posters manually and satisfying themselves they were real human beans before letting them loose on open forum.

Mods declined to expand on how they were checking for fear of the bots working out how to buck the checks!

I suspect a human registers the new account before letting their bot loose on it.

[XFool]

On another forum I use, there was a bit of a kerfuffle over regular users developing a habit of calling out new posters making generic bot-like posts, as bots. Team Mod asked us to stop it as they were vetting all new posters manually and satisfying themselves they were real human beans before letting them loose on open forum.

Oh no! The Return of the Pod People?

[UncleEbenezer]

I've seen a few posts on here now that are reckoned to be generated by 'bots'. How does a bot operate and how do they generate content. A user has to sign up and confirm membership via an email address etc. I'm not having a pop at this forum but there must be some kind of anti-bot technology to stop them from infiltrating excellent forums such as TLF. Perhaps the signing up process needs to be more stringent - 2FA maybe. Or do the bots have mobile phones as well

2FA doesn't help. It serves a different purpose: to verify the identity of an existing customer. I suspect it's not difficult for a bot to sign up with - generically - a PHPBB-driven forum. PHPBB or third-parties very probably offer anti-bot plugins, but none of them will be foolproof nor free of their own downsides.

I suspect a human registers the new account before letting their bot loose on it.

It would be for the mods to communicate with a new user and make a judgement of the replies - which could be quite offputting to perfectly legitimate new users and deny the community the benefit of renewal (as well as being an additional burden on the mods). It would have to be handled very carefully!

You could test that: register a new account and see what hoops you do or don't get to jump through. But I'd just ask the mods for comment before going behind their backs.

[gryffron]

A bot is a computer AI pretending to be a person. Yes, we suspect a real person signs them up an account then lets the bot loose to test if it can pass as a real person.

Looking back at our logs, about 1/50 users are quickly deleted as blatant spammers. We get c10 new users/day. We have had only a few potential bots, most of which seem to post a handful of times then give up of their own accord, probably after their owners realise they are failing the test. So bots are much less trouble than human spammers, and neither is unmanageable, so far. Our users are quick to call them out. Though we'd prefer you report their posts rather than calling them out publicly, as one suspect proved to be a legit person with poor English. So no plans to make the signup anymore difficult - yet.

Gryffron (Moderator)

[Mike4]

A bot is a computer AI pretending to be a person. Yes, we suspect a real person signs them up an account then lets the bot loose to test if it can pass as a real person.

Looking back at our logs, about 1/50 users are quickly deleted as blatant spammers. We get c10 new users/day. We have had only a few potential bots, most of which seem to post a handful of times then give up of their own accord, probably after their owners realise they are failing the test. So bots are much less trouble than human spammers, and neither is unmanageable, so far. Our users are quick to call them out. Though we'd prefer you report their posts rather than calling them out publicly, as one suspect proved to be a legit person with poor English. So no plans to make the signup anymore difficult - yet.

Gryffron (Moderator)

Yes I had this on another forum. I called out an obviously AI bot's first post as a bot and got an equally obviously human response, full of indignation and hot denial they were a bot! I still reckon it was the bot owner responding defending their bot but it shows the difficulties of calling out a suspected bot in the thread. That forum now requests members to report suspected bot posts rather than call them out out in the thread.

The T&Cs of that forum now have a new condition (something like) "No AI generated content allowed". But I'm not sure how Team Mod can tell. Team Mod says it can however, but are not saying how!

[stewamax]

Bots have been around since the Flood.
One evening in Xmas week in the early 2000s, I was running a phpBB board that had a blitz of bots trying to register. At its height between 20.00-21.00 we had 43 apparently different attempts (in addition to numerous retries). As Moderator Approval was switched on, we simply eyeballed the attempts, marked the apparently genuine ones (not easy since they were not actually posting anything at that point), and ran a PHP script to remove the rest. All before the days of reCAPTCHA et al.

[Mike4]

Bots have been around since the Flood..

You mean those we had back in February?

[UncleEbenezer]

A bot is a computer AI pretending to be a person.

ObPedant: a bot can be a whole lot of other things: for example, googlebot, which aims to crawl every website except where specifically instructed to keep out. I've written a few in my time, but not any that would register with or post to a forum like this.

[MuddyBoots]

Does anyone know what the point of bots are in forums like ours? Is it to just troll and disrupt the smooth running of the forum, or to actually try and influence us humans in some direction, eg for marketing or political persuasion?

[bungeejumper]

Does anyone know what the point of bots are in forums like ours? Is it to just troll and disrupt the smooth running of the forum, or to actually try and influence us humans in some direction, eg for marketing or political persuasion?

I imagine the point is to find out what passes the Turing test, and what doesn't? Once these originators are armed with that information, the highway to AI hell will be wide open.

I've said this a few times already, but the quickest way to spot a bot is to google its name. If you find that there are 33,000 other respondents of that name on other forums, all of which have signed up in the last week, then the chances are that you're onto a bad'un.....

BJ

[stevensfo]

On another forum I use, there was a bit of a kerfuffle over regular users developing a habit of calling out new posters making generic bot-like posts, as bots. Team Mod asked us to stop it as they were vetting all new posters manually and satisfying themselves they were real human beans before letting them loose on open forum.

Oh no! The Return of the Pod People?

What was the name of that film? I know which film you mean and still remember the original B/W film. We were staying with our grandparents and one of my younger sisters came down from bed in tears, saying that her younger sister was a pod!

No idea why they let them watch the film. Probably didn't realise how scary it would be,

Steve

PS Like a lot of the old B/W films, they relied on atmosphere and tension to scare us, rather than special effects. I seem to remember that it worked very well!

[XFool]

Oh no! The Return of the Pod People?

What was the name of that film? I know which film you mean and still remember the original B/W film. We were staying with our grandparents and one of my younger sisters came down from bed in tears, saying that her younger sister was a pod!

Invasion of the Body Snatchers

[UncleEbenezer]

Does anyone know what the point of bots are in forums like ours? Is it to just troll and disrupt the smooth running of the forum, or to actually try and influence us humans in some direction, eg for marketing or political persuasion?

I imagine the point is to find out what passes the Turing test, and what doesn't? Once these originators are armed with that information, the highway to AI hell will be wide open.
BJ

A lot of it will be perfectly innocent student-level pranks. Let's play! Bet you my bot can do better than your bot

At that level, the obvious ones are the losers. One up from that, they pass unremarked, like the one in the hummus thread. Winners will be those that draw serious human responses.

[Mike4]

Just wondering, do all these bots posts actually break any rules? If they do, which rule(s) please?

We human beans seem to instinctively feel they are an emerging problem but if they don't break any rules, should we not either accept them or change the rules?

[GrahamPlatt]

Just wondering, do all these bots posts actually break any rules? If they do, which rule(s) please?

We human beans seem to instinctively feel they are an emerging problem but if they don't break any rules, should we not either accept them or change the rules?

As UncleEbeneezer said, there are many types of “bot”. So let’s just concern ourselves with chatbots. Very clever mechanisms behind them, LLMs, but not intelligent. Do you want to be corresponding with such an entity, and accord it the right to be here?
Well, I don’t. Here, take the time to watch this please: https://yt.artemislena.eu/watch?v=NVxcsekcbhs

[MuddyBoots]

Just wondering, do all these bots posts actually break any rules? If they do, which rule(s) please?

We human beans seem to instinctively feel they are an emerging problem but if they don't break any rules, should we not either accept them or change the rules?

As UncleEbeneezer said, there are many types of “bot”. So let’s just concern ourselves with chatbots. Very clever mechanisms behind them, LLMs, but not intelligent. Do you want to be corresponding with such an entity, and accord it the right to be here?
Well, I don’t. Here, take the time to watch this please: https://yt.artemislena.eu/watch?v=NVxcsekcbhs

I correspond with an AI chatbot called you.com when I have specific questions as I find it provides well structured answers compared with a traditional search engine. You can ask follow-up questions without repeating the original one, and the answers are generally relevant to the question without trawling through many results from a search engine to decide which to use. Which is an issue of trust, because the chatbot is choosing the sources rather than me.

[UncleEbenezer]

Just wondering, do all these bots posts actually break any rules? If they do, which rule(s) please?

We human beans seem to instinctively feel they are an emerging problem but if they don't break any rules, should we not either accept them or change the rules?

As UncleEbeneezer said, there are many types of “bot”. So let’s just concern ourselves with chatbots. Very clever mechanisms behind them, LLMs, but not intelligent. Do you want to be corresponding with such an entity, and accord it the right to be here?

Bots have been doing great jobs for many years. And have been doing evil for almost as long. One with which I've been peripherally involved has passed a Turing test - in that she's been taken for human in interactive chat - many times over more than twenty years.

Come to think of it, to the question "what is a bot", the simple answer is, a robot. Geeks tend to be lazy, so we started using a single-syllable contraction of the word.

[csearle]

Just wondering, do all these bots posts actually break any rules? If they do, which rule(s) please?

We human beans seem to instinctively feel they are an emerging problem but if they don't break any rules, should we not either accept them or change the rules?

Not totally sure. If we decide we don't want them then, as you say, the rules can be easily changed. The subject is under discussion amongst the moderators and our admins.

My gut feeling is that it is a problem because, well at the moment at least, the content they produce is first-paragraph-of-Wikipedia kind of stuff. That
starts, by degrees, turning TLF into a boring search engine. Maybe if the AI technology was a magnitude better such that the answers it produced were so wonderfully insightful that few humans could have arrived at them then they might be beneficial. I still feel they should identify themselves as AIs though rather than masquerade as humans.

Challenge questions and non sequiturs seem to be the current way of getting a handle on their humanity (or otherwise). I suggest for the time being any suspicious posts are simply reported with an appropriate comment so we can take a view, probably by trying our own interaction. At the moment such reports are sufficiently few that the work involved is minimal. That could change of course.

We don't want to start deleting accounts because posters are sometimes weird in the way they express themselves - we'd all get deleted!

Chris
PS If TLF is being used for training AIs then they should bloody well pay for that service by agreeing a fee with the admins of this site.

[BullDog]

Much more of a problem if a number of bots start threads and have conversations amongst themselves without any human interaction? That's the logical end point, is it not?

For now, the posts I've seen here and reported are pretty obviously not human and rather poor quality. But that's going to change for sure.