[chrislehrer]

One of the best things about this site for many years now is the way people engage directly with others and their questions, pro and home. Lately I'm seeing a lot of what seem to be corporate posts that amount to "the problem you've described is a thing! Don't be afraid to experiment. Rah."

Far as I'm concerned, if you have nothing to say, I wish you'd not post.

 

[brianshaw]

Seconded. ChefTalk seems to have slowed down to nearly a crawl in the past year or so which has me always hoping for more participation. But the empty posts and what I suspect are AI-generated posts has me concerned/annoyed also. Not sure that much can really be done about it so at least we got to vent. Hopefully this discussion will inspire more participation with informational and responsive posts.

 

[butzy]

Hear hear!

 

[Iceman]

.


Listen ... All three a youse guys are all just describin’ a thing. Don’t be afraid to you know ... experiment. ... You’ll be fine.

 

[JustinJames]

They prefer genuine, helpful responses from individuals rather than superficial or promotional contributions.

 

[brianshaw]

This site seems to have been taken over by AI bots. Not worth my time any more, sadly. Time for me to take a break.

 

[chefpeon]

I have to say this AI Bot thing has me a bit flummoxed as a moderator. I've been wrestling with approving posts like these, because I know in my gut it's spam of some sort. I brought up the topic with the other moderators to get their thoughts on what I see as a growing problem. I couldn't get a whole lot of response there, which is unfortunate.

One the one hand, it's obvious those things are written by AI based on syntax and content. On the other hand, the responses are on-topic and basically correct. But I also see they seem to repeat what has already been said upthread. In those cases, I delete the comment but I don't ban the user.

I agree that a lot of pros use ChefTalk and look for genuine peer-to-peer advice and conversation. The AI generated posts really annoy me as a user, and the more I see of them, the better I am at spotting them. I had wondered if users here were as annoyed as I am. I moderate with a sense of fairness, trying to give each post the benefit of the doubt, as I really wasn't sure what the end game was for these AI Bot spammers. Regardless of the end game, I agree it makes this forum less appealing.

This thread has been helpful in telling me what I need to do going forward. Just trust my gut, it's spam, I know it when I see it, and I will be bringing the ban hammer down a lot more than I have been. Thank you very much for this feedback.

 

[brianshaw]

The dilemma is quite a dilemma!

Have you considered a captcha logon requirement like they do on other sites. Presumably that would be a verification of human, although yesterday I saw one of our suspected AI bots got through that. That kind of login is a real inconvenience but it might help mitigate moderation workload.

 

[ItsMeAgain]

XenForo has a "Spam Clean" option.
one click - bans user and deletes their posts.

superior solution to captcha, which btw is now about 80% vulnerable to bot cracking.

 

[chefpeon]

XenForo has a "Spam Clean" option.
one click - bans user and deletes their posts.

superior solution to captcha, which btw is now about 80% vulnerable to bot cracking.

We do have a "spam clean" option. I use it a lot. Not sure if it's from XenForo but it works well.

 

[butzy]

What about an initial 5 moderated posts?
I've seen that at a number of fora

 

[chrislehrer]

What about an initial 5 moderated posts?
I've seen that at a number of fora

Good idea!

 

[chefpeon]

What about an initial 5 moderated posts?
I've seen that at a number of fora

We have that in place. In fact, I think it's more than 5 posts. What really cracked me up this morning was there was something in the moderation queue that was posted to this very thread........it was an AI written explanation on how to deal with spam in public forums! I applied the ban hammer but good on that account.

I think the problem, at least with me personally, is that I really didn't pick up on the AI stuff until just recently. I initially thought, "wow, people are posting some really detailed content here". I think what should have clued me in was the formatting, like the use of bold fonts, indented paragraphs, bullet points.......all proper stuff, but nobody goes to the trouble to format like that in a forum. I should have known!

So you can blame me. I approved all that garbage in the first place. Now I'm sure I know AI when I see it and I promise you'll see less of that polluting this forum.

 

[brianshaw]

I doubt anyone here, except AI, will blame you, chefpeon. It’s subtle and insidious. Even detailed scrutiny seems to leave a shadow of a doubt and i actually applaud you for giving that benefit of the doubt whenever it seems appropriate.

 

[Iceman]

Let me just try and help out here just a little bit ...

NOT SPAM:

 

[ItsMeAgain]

Dear Ice,
and et.al. . . .

the "training" aka "education" of LLM (large language models) includes turning "AI" bots loose on forums, like here.
their aim is to 'teach' the program how to response, like a human, to a wide variety of "posts"

when the bot hits a (logical) dead end, the bot simply regurgitate the lines / quotes of the original message.

so, get Thee not upset "human spammers" are seriously easier to spot that AI bot who will post 10-20-30-40-50 messages before the include a link. remember - bot are fully automated, no people time required.....

 

[chefpeon]

Dear Ice,
and et.al. . . .

the "training" aka "education" of LLM (large language models) includes turning "AI" bots loose on forums, like here.
their aim is to 'teach' the program how to response, like a human, to a wide variety of "posts"

when the bot hits a (logical) dead end, the bot simply regurgitate the lines / quotes of the original message.

so, get Thee not upset "human spammers" are seriously easier to spot that AI bot who will post 10-20-30-40-50 messages before the include a link. remember - bot are fully automated, no people time required.....

A-HA! I knew there had to be and endgame/reason this was happening. Thanks so much for the enlightenment!

 

[counterprize]

they're back...

appear to be a new wave of single or minimal posters, including in the introduce yourself with suspiciously similar introductions.

 

[chefpeon]

they're back...

appear to be a new wave of single or minimal posters, including in the introduce yourself with suspiciously similar introductions.

Yes. I replied to your other comment in another forum about this. I'm sure you'll see it. But I'll copy it here because it's relevant:

I used to be a moderator here. After a while it became easier and easier for me to spot posts that were created with AI. There are so many different clues, like what you mention here; a fairly redundant comment resurrecting an old post. Other clues are formatting, cadence, content, etc. I would delete those posts without a thought. I'm fairly certain that those who read ChefTalk want real world answers from real world people, not AI or bots. I finally decided to leave as a moderator because it seemed non of my colleagues were on the same page or even cared.

Since then, I have seen many AI generated posts go through despite their obviousness. I messaged a moderator saying that AI undermines the ChefTalk experience and literally no one want to read AI content here. Their response was that if the content does not violate community guidelines then it's allowed through. I do not agree with this and I can see that AI is already making this site less useful to the community at large.

 

[chrislehrer]

Not only less useful but less interesting to participate in. To put it another way, this is killing the site. That means less ad revenue for whoever gets that from the site. I mention that point because clearly they don't care about the experience quality, but maybe they care about the bottom line? Here's hoping.