AI & Forums

mid_gen

mid_gen

mid_gen

Soldato
Joined
20 Dec 2004
Posts
18,257
So we have @Quartz asking recently if we should add AI functionality to the OCUK Forums, which was a resounding no.

But what of users just reposting AI slop verbatim? One of the great things about this forum, is that it is free from algorithmic influence, you just get the unvarnished voice of actual real people. It would be really sad to see this place reduced to just regurgitated AI nonsense, it's bad enough trying to identify the real posts amidst the sea of slop on Reddit, and the rest of the social media-verse.

I'll caveat this by saying that I use AI tools all day, every day, they're incredibly useful. But they can be dangerous. They are great for finding objective information, technical stuff, but people are using them to ask for opinions on subjective things....not understanding that their opinions are not trustworthy, if something isn't in their training data, it's just hallucinated nonsense.

If I were to suggest a policy, it would be 'no unattributed AI posts', and a total ban on AI responses in GD and SC.

KEEP OCUK HUMAN :km:

THANK YOU FOR YOUR ATTENTION TO THIS MATTER.
 
Currently I only see AI posts on here when someone is trying to back up their own talking points in a argument.

I don't mind if they are posting a snippet of 'troubleshooting' they've been given.
 
Make a proper rating system to sort posts by user rating and then any prospective ai slop would get sorted to the bottom. Or if ai posts are better, they'll get sorted to the top. Pure meritocracy, may the best post win. And/or impliment some ai detection and force a warning banner on any post suspected of being ai, and give them a forced pink theme. These are serious suggestions.

:edit: They were, in fact, not serious suggestions.
 
Last edited:
I agree it's time for an AI policy.
I don't like seeing copy-pasted AI responses on the forum.
I don't want to see AI images or video.
I'm concerned about the wording "no unattributed AI posts". I suspect I just don't understand what that means. Clarification needed on what is expected of users and moderators. This sounds like "attributed AI posts" would be permitted, which I don't agree with.
I would require that the policy is consistent for all forum sections.
 
throwaway4372 said:
I agree it's time for an AI policy.
I don't like seeing copy-pasted AI responses on the forum.
I don't want to see AI images or video.
I'm concerned about the wording "no unattributed AI posts". I suspect I just don't understand what that means. Clarification needed on what is expected of users and moderators. This sounds like "attributed AI posts" would be permitted, which I don't agree with.
I would require that the policy is consistent for all forum sections.
Click to expand...
I think in a technical forum, providing answers that may have come from a query to an LLM are ok? I mean I'd rather it didn't and the value of a real forum comes from people with actual experience of the problem at hand, but still, if it's providing solutions and links to useful info.

It's seeing the 'voice' of AI in posts that needs to be killed with fire. Responses in political threads that are just 'this is what AI thinks'.....no you fool, it doesn't *think* anything, it's just regurgitating a statistically plausible sounding series of words based on it's training data....of no value to anyone whatsoever.
 
The amount of AI is hard to monitor and its not going to any easier as time goes on. As the whole point of AI to mimic humans, walk, talk and think. Giving the posts people put on here.

All the anti push is going to make big tech be more strategic about how they push it. Like the constant push with blocking ads on from our devices.

If someone wants to prove their point then you cant stop them using AI to back their case. No different if they want to scan part of a book, upload it here and say "Look this is my point from xyz book I found"
 
Last edited:
Half the problem is that most people don't know how to use AI. Many are also using free ones that put minimal effort into "thinking", which leads to half-baked or just plainly wrong answers.

Tbh I don't see as much AI text here as I do on places like Reddit, but probably depends on the topics you're active in. It's a challenging one to solve because - if you ask people to be transparent, that'll be inconsistent and doesn't really solve the quality issue. If OCUK decide to monitor, there's a cost associated with that and also, what do they do when they detect AI content? You'd have to categorise it, which is another cost, or just blindly flag it, which isn't that helpful.
 
Mesai said:
Half the problem is that most people don't know how to use AI. Many are also using free ones that put minimal effort into "thinking", which leads to half-baked or just plainly wrong answers.
Click to expand...

We had this for years wayyyyyyyyyy.....before AI search was an thing. :D
 
Last edited:
I must be honest, the willpower required to resist posting up a ChatGPT response is quite substantial.

I do agree. I post up AI nonsense for the memes mostly but prefer actual thought out posts.

Dunno what we're gonna do about the jpaul situation though, when last did he have a firmware upgrade?
 
Diddums said:
I must be honest, the willpower required to resist posting up a ChatGPT response is quite substantial.

I do agree. I post up AI nonsense for the memes mostly but prefer actual thought out posts.

Dunno what we're gonna do about the jpaul situation though, when last did he have a firmware upgrade?
Click to expand...

Have you got that neurotypical bug?
 
Diddums said:
Dunno what we're gonna do about the jpaul situation though, when last did he have a firmware upgrade?
Click to expand...
It's called lateral thinking, as I replied to you, & you reacted to, in my recent post - aka thinking outside the box (sic)

I'm not sure people are able to distinguish sentences wordsmithed by AI now, also manifested in bland predictable journalism,
insidious use in amazon reviews, or reddit (as said), GenZ cv's ?
Livefree speech is the last bastion.
 
Sometimes I put effort in my posts when I'm particularly passionate about something but more often than not I'm on my phone browsing the forums and just don't have the time or effort to engage in every discussion I find interesting so will make a small / simple post just to contribute or add a little flavour to the discussion.

It takes a lot of effort to do long posts and it's enjoyable but ultimately it's just for a public forum and seems a little bit thankless putting a shed load of effort which may or may not be appreciated.

It's easy enough to spot AI generated posts but also it could just be someone putting in more effort for a particular topic.

LLMs are just predictive engines they don't think or argue or have knowledge but I guess you can be clever and merge its output with your own personality or perspectives?
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom