Published on 12, July, 2020
So, I'm slightly less grumpy than yesterday, as it's cooler.
Nevertheless, something else is bothering me, quite a lot.
Bots.
I know others of our Number are particularly concerned with this too.
Some bots are easy to spot as they leave links.
Others, make an innocuous post without a link but then later another post with a dubious link.
However ..... I've noticed another trend where a 'person' joins the forum to just post the once on a thread.
The post usually jars with me because although it sounds helpful, it doesn't sound to me like it comes from a real person and not an autistic person either.
The responses tend to me what I'd describe as 'glib'.
Question 1: Has anyone else noticed this?
It's happened on threads the last couple of days.
Do you think it could be bots too even though the response sounds relevant?
Also, I've occasionally wondered whether a 'person' is a bot even though they interact so I have another question.
Question 2: Is AI so well advanced now that bots can imitate humans in the way they respond on a thread and interact?
ie. they are not just posting in a vacuum but appear to be relating to you directly and your own responses.
Number said:the intent has been wholly nefarious
Quite possibly. I note a frequency of 'first posts' which I would hazard a guess are AI-generated. Often they are, in fact, requesting personal information under the umbrella (guise?) of shared experience.
Maybe these are some sort of harvesting bot, collecting data from which to learn.
A reply to one of your threads, Debbie, sounded straight out of the mouth of Chat GPT, but I didn’t want to say anything in case they are a genuine person.
Pegg said:Of course another problem is that humans can use Chatgtp to write text for them and there may be various reasons for it, not all of them nefarious.
I can confirm that this does definitely happen here, and that on a couple of occasions, the intent has been wholly nefarious.
Thank you both.
I'm going to check out Chat GTP.
Pegg said:nefarious
My word of the day
PS - question 2 - yes! Just try chat GTP. It is scary!
I agree, and there doesn't just seem to be one type of bot from one source, but multiple bot creators all with their own methods and agendas. Some agendas are obvious, like the ones that give links. Others, I just don't understand the motivation unless it's research into better bots. They seem to have got cleverer in the short time that I have been here.
I accept that I am of that age where I am more likely to be a luddite than embrace new technology, but AI to me just seems to be polluting everything.
The quality will get worse because soon they will be scraping the web and getting less human content and more of their own AI content and it will decay and decay.
Clickbait articles annoy the hell out of me too. Sometimes I am tempted to click and the article contains no actual content other than vaguely related words.
Debbie said:Question 2: Is AI so well advanced now that bots can imitate humans in the way they respond on a thread and interact?
I think yes, probably. Or at least that's the net effect, to an extent. I think a generic response, particularly in the language used and the sentence construction, can lead me to suspect AI, but even so, it probably won't be long before there's nothing to distinguish an online bot from an online human.
I agree with you re. the post you linked BTW, it reads like Chatgtp.
Of course another problem is that humans can use Chatgtp to write text for them and there may be various reasons for it, not all of them nefarious.
.