Forum menu
I don't understand such things, ref: tony-teacake
Influencing, casting doubt, sucking people in to adverse points of view, destabilising... etc.
Ah, a double bluff variant bot…
I don't get it either. No idea how it works or what the point is really.
I don’t understand such things
As I understand it, there are posts and whole thread(s?) that are "written" by AI 'bots'. On the surface vaguely believable, but with a few clues.
Why? No idea. I presume some kind of world domination / hacker / need links / spam.
I'm selling a car at the moment, and I note that most of the emails and texts are auto / bot messages now.
I'm sure there's going to be more of it.
I suspect the point is to see how well people are fooled really. This one always seemed to boil down to
1. validate people by mentioning their name, 'friend', saying 'good post' or referencing something said in a previous post
2. jam in a block of text from a relevant site
3. ask an open question to keep things going
A bit like a troll but cheaper and less lonely spaffing
Avoid Twitter and Facebook if you don’t get it. Much easier to clean up forums than it is social media.
I was about to post a similar question on the thread. What is the point, and unless something similar is posted like a bazillion times across loads of different platforms there isn’t a chance in hell that any population group are going to actually have their way of thinking altered. Especially when it mostly gets shot down in flames.
What is the point, and unless something similar is posted like a bazillion times across loads of different platforms there isn’t a chance in hell that any population group are going to actually have their way of thinking altered
It's not about people and opinion.
So what is the aim then?
Foreign actors influencing public opinion for their own political gain.
... or what the point is really
It's not always going to be obvious, but when carried out over many years, or even decades, it could have a significant effect. See Donald Trump. Of course, it's largely unquantifiable and very difficult to prove the actual impact, which is one of the reasons it's such an effective tool.
So what is the aim then?
On this forum it might just be to see how good it is getting. It was a good few posts in before it started to smell. Tweak the algorithm, repeat, see how far you get
"Yes, you're right about [insert keywords from previous post].
This reinforces my original point, which I'll repeat for you now."
AI bot chat is very similar to the way many politicians speak. Perhaps they receive the same media training?
Hmm, I’d do it to build a reputation for the bot that tries to establish itself a real user, as its doing human activity.
I’d do this for a lot of bots so I have a load of fake ‘real user’accounts, then I’d start using these to give likes on my political views/shopping opinions which would then push these opinions further up into the equations which decide what is shown to you on your feeds 🙂
Look up GPT-3, it's an open AI that people can use themselves to build these things. A couple of years ago one posted on Reddit for a week before people noticed.
https://en.wikipedia.org/wiki/GPT-3
That's an interesting point leffeboy, what do you think of the future for smelling bots?
On this forum it might just be to see how good it is getting. It was a good few posts in before it started to smell. Tweak the algorithm, repeat, see how far you get
Sort of like a Turing Test.
So just to confirm - we are working on the assumption that TT is 100% a computer program with no human involvement using an algorithm and access to Google in order to create the post and then answer the replies?
If this is the case I’m moderately impressed! How long before it becomes self aware?😵💫
Chewkw was v1, seems to have an upgrade to v2 recently, this was maybe v3…
Sort of like a Turing Test
The thing is, to pass that here all you'd need to do is post on one of the politics threads.
You could just repeat the same post for hundreds of pages with nothing more than a slight change to the wording and we'd ask assume it was absolutely genuine.
That’s an interesting point leffeboy, what do you think of the future for smelling bots?
IMHO it’ll probably be the having to lose the anonymity of registering with just an email and password and provide real world verifiable ID to prove your a real person.
You could just repeat the same post for hundreds of pages with nothing more than a slight change to the wording and we’d ask assume it was absolutely genuine.
+1
Ahh but that’s the catch affirmation an reiteration is a human trait 🙂
Sort of like a Turing Test
I loved that one of the factors in calling out Teacake was that his posts were TOO well written 😂
Possibly unfair to cite the politicisation threads but if you look at something trending on Facebook with a million comments, would you even bother to comment considering that it’s unlikely that anyones really reading the comments,as it would probably take a fair few days but people do(still post comments).
The thread in question, is it the Two stubborn people create a ridiculous moment ? 😉
I think you're a respectful metre behind on this sirromj
What? Posting without having read the preceding comments? On STW? My gosh, the end times are upon us! Save us Tony_T.
It was a good few posts in before it started to smell.
It stank from when they started the thread.
^ exactly.
It's making me realise that growing amounts on social media is now influenced by bots. Either direct posts or people that have been influenced by them. From spammy FB marketplace posts, to political views.
I *think* that some of the long term aim is about having as wide an influence as possible, across as many platforms as possible. This then can be used to either slowly steer opinion over time, or use all platforms at once in a 'thunderclap' moment of false news.
I guess they are also checking the site security and sign up system - so again one day perhaps brute force or hacking can take place?
IANAISE
It stank from when they started the thread
Well, for me it did, but some people were still responding when it was clearly a bot & other people were saying so.
Hmm, I’d do it to build a reputation for the bot that tries to establish itself a real user, as its doing human activity.
I’d do this for a lot of bots so I have a load of fake ‘real user’accounts, then I’d start using these to give likes on my political views/shopping opinions which would then push these opinions further up into the equations which decide what is shown to you on your feeds
Makes me glad that this place is my only online social activity. Wouldn’t the above just work for the feeds the bots were using across the various accounts? Sorry, bit thick when it comes to how this works and I’m now wondering which forumites might actually be bots!
Yep and gathering data on people to target them in the future for bog standard marketing or nefarious political mechanations.
Image of funkmasterp earlier
Wouldn’t the above just work for the feeds the bots were using across the various accounts?
I don't have the figures, but I'm pretty sure the number of fake accounts across the main social media platforms have been estimated in the billions. I think it's easy to underestimate the scale of influence, which at some point trickles into every day life offline, as people repeat what they've learnt online.
Whether it's political, commercial, or other, the actual AI side of things is a fast moving technology, making the the future a bit scary. It could quickly become very hard to determine what's real and what isn't.
So, doing a bit of digging I found Mr Teacake has a hundred odd posts over on the MSE forum, mostly posting within the house Buying, Selling and Renting board.
Bot or not, the entire thing stank. I appreciate this place does come up in google searches but is that more to do with algorithm providing confirmation bias or actual top hits?
. I think it’s easy to underestimate the scale of influence, which at some point trickles into every day life offline, as people repeat what they’ve learnt online.
Yep , I think the Cambridge Analytica stuff sorta pushed this out into the open.
It’s eye opening stuff.
It stank from when they started the thread.
To some people but not for everyone. That doesn't mean that they were more gullible, just that they were more willing to give the benefit of the doubt and have a discussion. At the moment these things seem to be more about keeping things going than influencing but I'm sure at some point this will change.
For example, yesterday I got what was obviously an automated call saying there had been dodgy activity on my ID and bank and before they issued an arrest warrant in my name I should press '1' on my phone to talk to the police. Obviously this was junk and once I pressed '1' I would get passed on to the human handler to do the final sale. With a bit of voice recognition and AI they might have been able to have more of a conversation and get more people to press 1 and increase their click through rate. I pressed 1 anyway as it was obvious and as I was walking to get a sandwich thought it would be fun to tie up the human for a bit.
To some people but not for everyone. That doesn’t mean that they were more gullible, just that they were more willing to give the benefit of the doubt and have a discussion.
Isn’t that the definition of gullible?
Or it was bot playing bot 😉
I officially don’t trust anything online anymore and think the OP is a bot. It started this thread as research in to how some guessed it was a bot 🤔
I pressed 1 anyway as it was obvious and as I was walking to get a sandwich thought it would be fun to tie up the human for a bit.
Unless of course pressing one just connects you to a premium rate number in Djibouti and the first you'll know about it is a 600 phone bill at month end. Or agrees to receive premium rate texts once a month etc.
Its not like these things are going to be entirely honest about what you're actually agreeing to with your button press is it?
To paraphrase what has been said on the scam bike sale thread, they're not as dumb as you think they are, you're probably exactly as dumb as they think you are.
Just stick to liking cat pics on insta and Facebook and you’ll be fine 🙂
To paraphrase what has been said on the scam bike sale thread, they’re not as dumb as you think they are, you’re probably exactly as dumb as they think you are.
Yep the only way to win the game is not to play.
I doubt in this case it was intended to achieve any aims - I think this one was purely using the forum to make the bot better. It's been a little while since I learnt about it so I'm probably talking bollocks, but my understanding is that you have to teach e.g. a neural network by feeding it information and scoring the output. So by replying you're giving it examples of the correct responses which in turn help it sound more like a real person.
Isn’t that the definition of gullible?
Hooked a few of the usual ‘I just want to debate’ crowd though didn’t it.
Makes you think.