The GPT pandemic

Oooh I started seeing this today.

This was in a thread about Bluetooth technology, checked it with an online 'checker'

I checked this 'checker' against other replies in the thread, and they were 90%+ being human as a reply.

I can see this being a problem to raise the 'reputation' of the account without being noticed by spam filters, and especially without an attentive userbase to report content like this.

Oooh I started seeing this today.

This was in a thread about Bluetooth technology, checked it with an online 'checker'

I checked this 'checker' against other replies in the thread, and they were 90%+ being human as a reply.

I can see this being a problem to raise the 'reputation' of the account without being noticed by spam filters, and especially without an attentive userbase to report content like this.

What online checker did you use Christopher? 

https://writer.com/ai-content-detector/ 

Thanks, that looks useful! We'll keep an eye out and see how it does. We have a chat on our community about ChatGPT and I've just used the tool to check something that I know is AI produced and it detected 5% human generated content. Interesting stuff! 

I've been doing a lot of research into ChatGPT and related tech, and unfortunately it's pretty easy to have ChatGPT generate content that will present itself as 100% human generated in those checkers, you just need to know how to provide the proper instructions.  I'm seeing people document lots of ways to do this. 

Some of these AI checkers have APIs for them.  It shouldn't be too hard to develop an automation that checks content when it is posted, and automatically flags it for moderation.  It won't be perfect, but maybe it can help a little.

I'll be a little more concerned if GPT's being used to fill out user profiles to make them indistinguishable from spambots, and then the content

Hooking into an API to check it is a good idea though, one of the many data points that can help determine the state of an account that's signed up.