Thoughts welcome: Rule against AI-generated posts?

[Second Summer]

I'm considering adding a rule against AI-generated posts.

Some reasons:

1. AI-generated content is prone to hallucinations, i.e. falsehoods or fake facts.
2. I don't believe members want to read AI-generated posts or answers to questions, even when they're factually correct, but rather they want authentic posts written by actual humans, even if the grammar or vocabulary is poor, questionable punctuation etc. If I wanted AI answers, then I would ask an AI directly. Which in most cases I won't.
3. AI-generated posts contribute to ever worsening information quality on the web. AIs trawl the web for info and use it as training data, then AI answers are generated based on this, some of which find their way back to the web, then that is again used as training data. You can see how this becomes a vicious cycle where hallucinations and falsehoods are amplified.

I do realise it will be difficult to enforce a rule like this, and no, it hasn't been a huge problem yet, but I have seen several obviously AI-generated posts.

 

[Lou]

I do realise it will be difficult to enforce a rule like this, and no, it hasn't been a huge problem yet, but I have seen several obviously AI-generated posts.

I don't have any issue with this rule at all but the thing about enforcing it...
I think I too have seen some VF AI posts. But I think since they Were obvious they were also harmless. Although it probably is a good idea to protect the gullible. And as AI get better the bar for obvious-ness is going to shift and create more Gullibles.

Could a non-native English speaker be mistaken for AI? Especially if he relies on a computer based translator. Some thought and care will need to be applied.

Will all VF members need to pass a test proving they are human.

Which then reminds me of this joke....

 

[Lou]

You can see how this becomes a vicious cycle where hallucinations and falsehoods are amplified.

Ha! We don't even need robots - People are pretty good at doing this to themselves. See: Haitans are eating cats.
(However, I have no problem if you ban Trump and Vance from the forum. )

 

[beancounter]

Yes, add the rule.

 

[Second Summer]

Could a non-native English speaker be mistaken for AI? Especially if he relies on a computer based translator. Some thought and care will need to be applied.

That's a good point which I'll try to keep in mind, thank you for mentioning that.

Will all VF members need to pass a test proving they are human.

Not existing members but maybe we need better screening of new members. For example, maybe it should be a requirement that every new member makes an intro thread before they're allowed to post elsewhere.

 

[Lou]

Not existing members but maybe we need better screening of new members. For example, maybe it should be a requirement that every new member makes an intro thread before they're allowed to post elsewhere.

Regardless of AI Bots, I like that idea. Or maybe a Captcha. or some other kind of question.

Although, is it really a problem? I maybe have seen one bot here. But maybe you get rid of them before I see them.

years ago in another forum we had a problem with bots wandering around than dropping advertisements or other spam. you must have solved that problem - something automatic?

 

[Tom L.]

It's up to you. I certainly wouldn't miss anything if AI-generated posts were excluded. I tend to quickly skip over such posts as it is. On the other hand, if blocking AI posts is more work for you, I can deal with the AI-generated spam. I appreciate having this website.