View Single Post
  #3  
Old December 8, 2022, 05:53 PM
GordonJ's Avatar
GordonJ GordonJ is offline
Administrator
 
Join Date: Aug 2006
Location: West Palm Beach, FL
Posts: 3,480
Default Re: The "risk" with AI-generated text is...

I see, you too, get the MJ DiMarco newsletter. So, I signed up, spent all of 15 minutes there, and left. Nothing there for me.

I'm so happy that I won't be around when HAL tells Dave (2001 A Space Odyssey) "don't touch that Dave".

We're not far from AI to AI, and then, what's the point. My bot will answer your bot...I want to set up my old answering machine, cause the robot calls I'm getting are ridic...and since I gave up may phone number at this site, I can't wait to start getting their messages telling me how great their AI is.

My creative writing
high school teacher would certainly have given me a D- for that story. And that just for turning it in on time.

I don't know how MJ used it, but more power to him.

Thanks for sharing, but now, what was the big secret?

Was it...."at least with AI, there is some proof there might be a little intelligence on the Internet" A very little.

Mr. (Exhales noxious Cohiba smoke into the air) Alexandre




Quote:
Originally Posted by Dien Rice View Post
The danger with using AI-writers (at least at the moment) is that it can often be detected...

If you put the above story into the AI-detector at

https://huggingface.co/openai-detector/

It detects the writing as being "99.98% fake"... (That is, written by AI, and not a real person...)

I don't know how it does it!

I have a feeling, though, as artificial intelligence gets better and better, it will be harder and harder to detect...

The other thing is that AI often gets its facts wrong - if you're trying to write a factual article! It "makes up" facts, which can also be a problem...!

- Dien

Last edited by Dien Rice : December 8, 2022 at 10:05 PM. Reason: fixed formatting
Reply With Quote