Can Humans Detect AI?

[getakash1]

So, I was wondering if humans can actually detect AI . Just for Fun, AI detectors are bad as we all know.

So, i wrote the following Piece, with either AI/myself, what do you think it is? I will update the thread with the answer tom.

------------------------

Hey BHW fam!

So, I've recently joined this crazy workshop on SEO, and it got me thinking...

Why doesn't BHW host workshops and conferences?

Imagine:

• Networking: Chatting face-to-face with fellow black hatters.
• Learning: Picking up new tricks of the trade.
• Collaborating: Finding partners for your next big venture.

Wouldn't that be rad?

Your thoughts?

Would you attend? Let's brainstorm!

 

[Vendettta]

Bro, "bhw fam" is a dead giveaway lmao

 

[Beklink]

100% AI

 

[lukey372]

Yes, anything that has "Myriad" in it is 100% written by AI. 99% of people don't know how to use the word properly and if you generate shit via AI 9/10 it'll pop "Myriad" in your article. It's a dead giveaway.

This isn't written by AI.

 

[profilinjays25]

So, I was wondering if humans can actually detect AI . Just for Fun, AI detectors are bad as we all know.

So, i wrote the following Piece, with either AI/myself, what do you think it is? I will update the thread with the answer tom.

------------------------

Hey BHW fam!

So, I've recently joined this crazy workshop on SEO, and it got me thinking...

Why doesn't BHW host workshops and conferences?

Imagine:

• Networking: Chatting face-to-face with fellow black hatters.
• Learning: Picking up new tricks of the trade.
• Collaborating: Finding partners for your next big venture.

Wouldn't that be rad?

Your thoughts?

Would you attend? Let's brainstorm!

Yes and No, there are various strong indicators in pieces of text depending on the context and use case that are red flags. Such as "BHW Fam!", this instantly alludes to the writing of something that doesnt know the context behind the community they are trying to communicate with, however at the same time this could just be a humans writing purposely trying to throw someone off. It is also possible to create better prompts for a better output, furthermore tweaking the prompt and pre-providing the AI important information such as the context behind what its writing.

 

[getakash1]

Yes and No, there are various strong indicators in pieces of text depending on the context and use case that are red flags. Such as "BHW Fam!", this instantly alludes to the writing of something that doesnt know the context behind the community they are trying to communicate with, however at the same time this could just be a humans writing purposely trying to throw someone off. It is also possible to create better prompts for a better output, furthermore tweaking the prompt and pre-providing the AI important information such as the context behind what its writing.

Wow, is this answer written by AI ? Hahaha , nice job man , must be a good prompt, almost sounds human.

 

[jlp]

Wow, is this answer written by AI ? Hahaha , nice job man , must be a good prompt, almost sounds human.

The way I do it to get such results is to give AI instructions of what I want, and then give the AI a sample of text, and ask to copy the style of the original writer. It is way better to ask him to imitate rather than explain the style you want.

 

[getakash1]

The intitial text i gave yesterday was 100% written with AI xD

Result : Humans can detect AI content with 80% accuracy



How can we expect 100% Accuracy from AI detectors? Food for thought.

 

[adlogik]

Lol easily!

 

[Project Limitless]

The intitial text i gave yesterday was 100% written with AI xD

Result : Humans can detect AI content with 80% accuracy

View attachment 290732

How can we expect 100% Accuracy from AI detectors? Food for thought.

The AI detectors are not reliable. There is no way to accurately detect AI content.

 

[seobond007]

The human can detect AI but not 100%.

 

[RATIKA]

Yes, Easily, I am using 4-5 apps to generate content and i notice the way of generating content is mostly same.

 

[itz_styx]

generic AI content can often be easily spotted by a few indicators when you've seen enough. it needs good prompts to avoid detection.
however humans will never be able to 100% spot AI content everytime and software AI detectors are mostly garbage anyways, they are too unreliable and also flag human content.
even OpenAI discontinued their own AI detector as it wasnt reliable enough, which says a lot

 

[espacefree]

The formatting pattern AI uses and the perfect grammar + punctuation gives it away, humans make much more variations ( eg. don't use commas perfectly every time, type space two times or do not type it when it's needed, use several writing styles/tones in the same article, etc.)

 

[MoonStarz]

I'll go for Human just because this "and it got me thinking..." three dots are used i know but never saw AI responses with that so that make me think is human not AI.

 

[JavascriptPufferfish]

its either written by AI or by someone who has seen enough AI writing to imitate it. I mean, who the fuck says "this crazy workshop on SEO". that sounds like you told chatGPT to play a "fellow kids" persona lol.

 

[leechesBleechin]

Ai detection is basically, look for random discussions about morals and ethics that don't belong in the context of the thing. That call to action at the bottom is all gpt threw and threw. The almost robotically perfect grammar and spelling and in general it just gives off an AI aura (if that makes sense?)

 

[TheoreticallyBr]

YES you can notice ai content based solely on the flow. it has no soul in it.

 

[Monzani]

In general, it's easy to identify IAs because they try to make disclaimers all the time to ensure safety and follow politically correct behavior.

In your text, this could be an IA, or it could also be a boring TV show host trying to be engaging in an awkward way...

 

[antho_rob]

Human can detect and also cannot detect Ai in some case, meaning humans cannot detect Ai 100%