AI infiltrating web forums?

[great white]

While I haven't seem much of it here, I am starting to see AI popping up on some other web forums I frequent. They're mostly car hobby web forums where it's pretty easy to spot a generic AI response to a highly specific or technical question.

AI posts are actually pretty easy to spot at this point. They usually start with a reply to a post with something like "It seems you're having a problem with...." or it's a wall of text with bulleted lists, with formatted indents spacing and all. The sort of post a human doesn't make on a web forum (unless they're an english major). Certainly not a format most "gear heads" type in responses.

Of course, my concern is that as AI learns how humans interact in these settings is that it will learn that perfect grammar and spelling is often the "tip off" that it's not a human but a machine responding and starts to incorporate those "errors" and gets harder to spot.

I often wonder if the moderators on this forum are perhaps watching for it so it doesn't show up and ruin the forum interactions between real humans as HM seems to be pretty free of these AI intrusions.

I know myself, once I see an AI response on a thread, I generally loose interest in the topic as I don't want to be talking/reading a machine response to a human question.

I feel AI has it's places, but not in human to human discussions.

I guess I'm mostly just rambling. I don't fear the coming of AI, it's just another tool/technology we developed. What I do worry about is how humans deploy and use it.

I guess I'm kinda looking at how social media came out as such a great thing and has been mutated into this horrible thing we've released on our society (most especially, our children). It's pretty easy to see AI going the same way: lots of promise and possibilities at the beginning and getting warped into something that ultimately is bad for us.

Anyone else have thoughts on it?

 

[markba633csi]

Hey, if it tells me how to fix miswired contactors on a Chinese lathe I'm all for it!
The deep dream generator is pretty fun, check it out

 

[woodchucker]

I guess I'm mostly just rambling. I don't fear the coming of AI, it's just another tool/technology we developed. What I do worry about is how humans deploy and use it.

I have no faith in humans. We always seem to drop to the lowest points. Thefts, murders, sales pitches... look at all the click bait ads, even in reputable news feeds.

Man always seems to (gravitate) no SINK to the lowest uses. To make money without working hard. To find a way to kill others without risking life (drones) (therefore no skin in the game).. No, I do not relish the coming of AI. I have watched people code back doors in their code for years. In scarry situations, put simple passwords into secure systems so they can get in without trouble (so can others) these were often the idiots, who were lazy and created the worst backdoor, the easy hack...

AI has potential to do good, but it has way more potential to be pandoras box.

 

[Alcap]

I been trying to follow and understand AI and on the surface seem very interesting and scary at the same time . Going though the years of reading magazines and books to having most of that information free on the internet I dont subscribe to and print though I still by books . I wonder how AI will change things going forward?

 

[higgite]

I'm just glad my name isn't Al.

Tom

 

[Mauzinette]

I’ve been studying the use of AI in bioinformatics for a good while and these so called language models are getting dumber, not smarter.

The problem in a nutshell is that AI is very poor at detecting AI created content and it has started to kind of deteriorate itself.

ChatGPT use declines as users complain about ‘dumber’ answers, and the reason might be AI’s biggest threat for the future

AI for the smart guy?

www.techradar.com

ChatGPT can get worse over time, Stanford study finds | Fortune

ChatGPT went from answering a simple math correctly 98% of the time to just 2%, over the course of a few months.

fortune.com

OpenAI Quietly Shuts Down Its AI Detection Tool - Decrypt

Dashing the hopes of liberal arts professors and teachers, OpenAI decommissions its AI content detector due to poor accuracy.

decrypt.co

If you’re still worried about AI, ask it to draw a dog in a machine shop





On the second try it even missed the dog

 

[Mauzinette]

While traditional machine learning also suffers from various drifts and biases, it is more manageable as the use cases tend to be quite definite.

Very useful stuff, saves time, but won’t take over your Internet forum

 

[woodchucker]

the ai that you guys are referring to is not the full AI.

There's been a lot of work that the DOD has been doing . IBM and others for years, Watson which has been doing medical and other problem solving.

The current AI that we see is relatively not smart.. it's record , digest, and spit out. It's danger is misinformation..

I am also worried about the misinformation. That is already very destructive to our country, that and all the conspiracy believers. Everything these days seems to be a conspiracy, even simple stuff. So it tears us apart.

 

[silence dogood]

AI or "artificial idiocy" is like a back seat driver from hell. I pops up and tells me what I want to chat about and then it even asks if I want to chat about other subjects that I have no interest in. There is no way that some computer program could possibly know how I think. So, I hit search in the upper left corner and do my own search.

 

[Eddyde]

Yes the moderators here at HM are on the lookout for such AI, Spam-Bots and ban them when found.
However, if a human member decides to use AI to generate their post, we don’t have much we can do about it.
If anyone suspects AI generated content, please report it and we’ll investigate it.
Thanks