Is there any solution to bypass originality ai?

NerdofIM

NerdofIM

Regular Member
Joined
Jun 3, 2018
Messages
317
Reaction score
108
The texts generated by gpt3/gpt4/bing etc are all being detected to be 100% AI written by originality. Asking gpt to passphrase the text did not solve it. quillbot passphrase deliver incoherent texts (plus it is not free) .

Do you know any working trick for bypassing AI detectors ?
 
Advertise on BHW
What do your prompts look like? What type of writing style are you telling the AI to write with?

Prompting correctly is important and can make undetectable AI content, but you should still edit each article a little bit afterwards.
 
NerdofIM said:
The texts generated by gpt3/gpt4/bing etc are all being detected to be 100% AI written by originality. Asking gpt to passphrase the text did not solve it. quillbot passphrase deliver incoherent texts (plus it is not free) .

Do you know any working trick for bypassing AI detectors ?
Click to expand...
which detector are you using?
 
Undetectable Ai claim to be able to pass it but I find good prompts work well. It’s also something not many people are really bothered by.

Good AI content is good content.
 
PHPInjected said:
What do your prompts look like? What type of writing style are you telling the AI to write with?
Click to expand...

I ask the AI to create a passphrase after delivering the initial text. Then, I ask it to write the text in a more human-like way. This decreases the percentage of AI-written text, but it’s still being easily detected. Do you have any tips for improving the prompts?
 
NerdofIM said:
originality ai and copyleaks



mind to provide some prompts?
Click to expand...
So I will often get chat GPT to rewrite an article and then provide specific instructions

Rewrite the following article in the style of xxxx include a related anecdote and personal examples to illustrate points. Address the reader directly. (Then paste part of the article I want it to rewrite).

Edit. Just checked. My article from my prompt was 57% human. Used undetectable and it was 99%.

Only tried it once but for those that care it’s encouraging
 
NerdofIM said:
Asking gpt to passphrase the text did not solve it. quillbot passphrase deliver incoherent texts (plus it is not free) .
Click to expand...
NerdofIM said:
I ask the AI to create a passphrase after delivering the initial text.
Click to expand...

The issue here might be the use of the word 'passphrase' - I think the word you mean is 'paraphrase'. Try that in your prompts instead.
 
it's all about prompting tbh. Plus you can add some content in your prompts and ask chatgpt to base the article on it or something.
 
AI detectors are complete garbage and are being accepted as a legitimate tool based solely on perception and not reality.

Test them out with ANY original content from a wide range of samples and you'll see there is no consistency.

Even part of the Bible came back as AI content in one test. Go figure!
 
PlinkPlinkFizz said:
AI detectors are complete garbage and are being accepted as a legitimate tool based solely on perception and not reality.

Test them out with ANY original content from a wide range of samples and you'll see there is no consistency.

Even part of the Bible came back as AI content in one test. Go figure!
Click to expand...
then the bible is written by AI? LOL
well, it's basically just a witch hunt, there is no way to accurately detect which one ai or real if it's of the same quality.
even if GPT watermark article, there are tons open sourced LLM models we can use locally.
 
PlinkPlinkFizz said:
AI detectors are complete garbage and are being accepted as a legitimate tool based solely on perception and not reality.

Test them out with ANY original content from a wide range of samples and you'll see there is no consistency.

Even part of the Bible came back as AI content in one test. Go figure!
Click to expand...
Exactly!

I really don't know why people still focus on AI detection when there are much more important metrics. Google already said that AI content is ok, so why waste any time trying to come up with non-AI content?
 
PlinkPlinkFizz said:
AI detectors are complete garbage and are being accepted as a legitimate tool based solely on perception and not reality.

Test them out with ANY original content from a wide range of samples and you'll see there is no consistency.

Even part of the Bible came back as AI content in one test. Go figure!
Click to expand...


I understand your point, AI detectors are not 100% accurate, however they are able to detect AI contents from hand written content in most cases. I tried it a few times and it did not flag a hand written text as AI by mistake.

V8200T5 said:
I really don't know why people still focus on AI detection when there are much more important metrics. Google already said that AI content is ok, so why waste any time trying to come up with non-AI content?
Click to expand...

The problem is that Google change the algo and the rules from time to time , and I think that it may outrank hand written content currently or in the close future, so I prefer to learn how to generate hand written content using AI tools
 
Can someone explain in simple terms how an AI detector works? Because based on the discussion above, I have doubts about its actual functionality.
 
In many instances, AI detectors can be inaccurate, and even after making edits, the output might still be labeled as 100% AI-generated.
 
Advertise on BHW
You must log in or register to reply here.
Sign up now!

Main Menu

Home Main Forum List Black Hat SEO White Hat SEO BHW Newbie Guide Blogging Black Hat Tools Social Media Downloads

Marketplace

Account Selling Content / Copywriting EBooks / Methods Hosting Misc Proxies For Sale SEO - Link building SEO - Packages Social Media Social Media - Panels Web Design More

Making Money

Affiliate Programs Hire a Freelancer IM Journeys Making Money Pay Per Click Site Flipping Want to Buy

BlackHat World

Dispute Resolution Forum Suggestions & Feedback Introductions Lounge
Back
Top