How to Bypass Originality.AI Detection

Disclaimer: The contents of this article are purely educational and aimed at fostering a deeper understanding of AI and plagiarism detection mechanisms. While we present strategies that could potentially bypass these systems, the intention is not to encourage dishonest practices like plagiarism. It is crucial to understand that using someone else's work and presenting it as your own is both unethical and can be illegal. The techniques discussed here are meant to highlight that AI detection tools can be quite faulty. Misuse of these techniques for unethical purposes is strictly discouraged and not supported by the author.

I've never been more excited to actually get stuff done then when ChatGPT came out. I get so much writing, organization, and editing done in a tenth of the time it would've taken me a year ago.

Even last summer when I was writing blogs every day I was legitimately spending hours by the keyboard typing subpar sentences with the hopes of growing a blog. And then... ChatGPT was released.

At first it was like a revolution took place and there was utter chaos and destruction in the streets. What was gonna happen to freelance writers, journalists, technical writers. Was everything gonna get replaced? Not really. But it did spawn the start of AI detection companies like Originality. If you want to learn how the entire tool works you should check out our full review.

While I like and use Originality quite often to check my very own writers work, it's really not perfect. Human-written stuff can be detected as AI while ChatGPT + Undetectable AI + HideMyAI often gets bypassed.

No matter how much you try to make your writing look human, it's not really up to you to determine whatever these detectors will classify it as. If you have human-written or ChatGPT-assisted writing that is getting constantly flagged as AI with Originality, here's what you should do to get around it.

Use AI Detection Bypassing Tools

The first thing if you don't want to spend much manual labor tweaking your writing is to use one of these bypassing tools.

Paste your writing into an AI detection bypass tool like Undetectable AI or HideMyAI and having it rewrite and change your content around.

All this detection/undetection stuff is like a dog chasing their own tail. AI bypassing tools get better and then detection tools get better, and this pattern just keeps going back and forth.

It doesn't help that OpenAI discontinued their own detector and even said AI detection tools don't work. While I think detection tools give decent hints, OpenAI is right, they don't "work"

To use one of the undetection tools in the meantime, you can simply ask ChatGPT for a paragraph & then paste it into a tool like Undetectable and you'll get a brand new revised & expanded paragraph.

You can change the readability level & purpose of writing. If you want to rewrite a ChatGPT article to be at a university level general writing, you can. If you want a PHD-level academic essay, you could also have it rewrite content for that.

You can change the setting from balanced to either more readable or more human. More readable will keep things similar to what you pasted, and more human might still get picked up by Originality. Again, this really is just a coin toss. I feel bad for innocent students who don't even use ChatGPT and still get in trouble.

Don't Rely on AI Completely

I use ChatGPT a lot, and I used to over-rely on it when it first came out. The more you use it the more you realize you can't rely on it to write and produce everything you need.

Some tips I follow when using AI to write now are:

• Have it draft an article outline
• Rewrite & improve an article I originally wrote
• Fill in the end of paragraphs when you hit writers block
• Brainstorm ideas back and forth
• Article proofreading & grammar checking
• Summarize research before adding personal touch

You really can only rely on ChatGPT for things that don't change. A great example of this is history. If you need a paragraph explaining why a certain war took place, go ahead and ask. That information isn't going to be different whether you write it yourself or ask the robot.

Where things do get personal is the opinions. You can have ChatGPT generate you the outline to a certain war you're writing about & give you some ideas for persuasive opinions and take it from there to keep writing.

Physically Edit Your Writing

The next thing you could do, assuming you asked ChatGPT for an entire article, is to go through it and edit all the annoying and robotic phrases it loves to use.

One of the easiest ways I know someone used ChatGPT to write something was when it uses language like "it is important to note," "it's essential to", "it should be noted," "it is worth mentioning," or "in conclusion." These phrases, while not inherently wrong, are often overused by AI and can give away the non-human nature of the writing.

I've also noticed a ton of "Moreover" and "Furthermore" sprinkled throughout the text.

These transition words, although useful in certain contexts, are often overused by AI, making an article sound more mechanical than conversational.

AI might also repeatedly use phrases like "significant impact," "a wide range of," or "essential aspect," which can be quite noticeable and repetitive. You need to tailor the output to more closely match your unique writing style and tone, not a textbook.

Spend a few minutes looking for these words and just make them look human. Remove things you wouldn't say in person. Aim to revise like you're trying to speak your article out loud. Would you say that in real life? Probably not... even in a formal setting.

Remove Repetition

Eradicate repetitive content. AI can sometimes fall into the trap of redundancies, repeating the same information in different ways, or using the same examples and phrases over and over again. It's so obvious too.

If a tool runs out of ideas or creative ways to express something, it'll just try to sum it back up or repeat a bunch of bloated garbage.

ChatGPT tends to love certain phrases or ways of explaining things, and it may lean on those too heavily. Scan your work for repeated words or phrases and substitute them with synonyms or rephrase the sentences. Consolidate your points where it makes sense or use a tool like QuillBot to paraphrase otherwise repetitive sentences.

Also, try to add some variety to sentence structures to avoid the monotony of same-length sentences. Combine short sentences or break up long ones, vary the rhythm, and include different types of sentences - statements, questions, exclamations - to keep your reader engaged.

Only Use ChatGPT for Outlines

I know I mentioned this earlier but the more I write the more I realize it makes sense. If you try to over-rely on ChatGPT you'll end up in a pit of complacency. You won't want to change anything because you'll just think: damn that's so good. But it's really not. Generally it's repetitive and the extremely formal writing style just doesn't make you sound like a normal person.

If you're a scholar trying to write PHD level work, you might be good out of the box. But you won't maintain your original writing style by having ChatGPT write most of your stuff.

Implement Paraphrasing & Replacement Techniques

The last thing you could do to bypass the detector is really just paraphrase your content. I've worked with QuillBot for quite a while to help do this and it works super well. You have a ton of customization options and levels of replacement depending on what you want.

You simply just paste your writing into QuillBot & let it do the rest:

You'll see a revised output that has a completely different result compared to what you put in. You can change the mode of output (to get more formal, creative, or simple) and change the rate at how many synonyms you want placed in the paragraph.

This sometimes does the trick, but again it's important to just use this as a tool to replace a few sentences, not replace your entire writing.

Final Thoughts

It's a tricky situation. You don't want your writing getting flagged if you just used ChatGPT for a few ideas and sentences, so you really have to get creative. AI is an amazing supplement to human writing, but I think we're all quickly learning it's not a complete replacement yet. Only time will tell where things will be in a few months from now.

You really have to just keep your writing authentic throughout the entire journey. Get tips from AI but don't over-rely on it as you don't want to get picked up by one of these tools. You're still a human after all of it, and bypassing Originality really just comes down to staying unique. If your writing gets flagged by Originality or another detection tools even though you're not even using AI, just know these tools are not provably accurate. They just give estimations.