HOW did this AI generated content bypass Ai detection tools?

Lenz

Registered Member
Joined
Aug 10, 2021
Messages
96
Reaction score
35
Hey,

I noticed a website with ai generated contents and I'm super sure it is written by Ai. Because I generated the similar content with the same tone and wordings. But that person's text goes "human written" on an Ai detection tool whereas mine goes " Ai detected" What is that other person doing that could bypass Ai detection very easily? Any secret tools or insights? Thanks!
 
I suppose he made a human touch after generating the content. This is totally possible.
ugh, like rewriting the whole content? but chances for it to be "human written" is not 100%
 
There are some tools that offer this, not 100% human written, though (and also depends a lot the tool you use to check)
 
ugh, like rewriting the whole content? but chances for it to be "human written" is not 100%
It seems tedious but think about the past when AI was not available. That we made huge content right? So, AI just makes the whole process faster and easier.
 
The other person might be using advanced content optimization techniques or tools to make AI-generated content appear more human-like. Consider refining your content generation process and fine-tuning the AI settings to achieve similar results.
 
Using advanced prompt with some manual edit can help bypass it.
 
The 'AI content detectors' are about as useful as a chocolate teapot. Simple stuff like limiting the length of syllables per word that the model will output, or running the generated text through a T5-powered paraphraser will usually throw them off. The info they give you is a placebo at best.

And do you really think Google will be running each page of your website's content, probably equalling at least a few thousand tokens each, through a LLM capable of anything useful?
 
thought about giving a human touch
By making some mistakes maybe in typing or in spacing etc.?
I am still looking for it
 
The 'AI content detectors' are about as useful as a chocolate teapot. Simple stuff like limiting the length of syllables per word that the model will output, or running the generated text through a T5-powered paraphraser will usually throw them off. The info they give you is a placebo at best.

And do you really think Google will be running each page of your website's content, probably equalling at least a few thousand tokens each, through a LLM capable of anything useful?

you're right about the chocolate teapot and other facts too ;)
 
Hey,

I noticed a website with ai generated contents and I'm super sure it is written by Ai. Because I generated the similar content with the same tone and wordings. But that person's text goes "human written" on an Ai detection tool whereas mine goes " Ai detected" What is that other person doing that could bypass Ai detection very easily? Any secret tools or insights? Thanks!
hello!
undetectable AI is a tool we designed to detect AI written text and make AI-generated content more human. We do this through a variety of machine learning models and additionally use text pre-processing and post-processing to ensure that AI detection has been removed and you are able to bypass AI detectors.
 
Back
Top