How to Bypass AI Writing Detection Tools to Humanize Your AI

To make writing appear more human, you have to write it like a human! This means more creativity, lack of predictability, and authenticity. Here's a list of our best tips to help bypass AI detection tools by humanizing AI-written content.

Justin Gluska

Updated January 11, 2024

a friendly and cute robot sitting in a lawn chair on his porch typing on his laptop as photorealistic 4k digital art

a friendly and cute robot sitting in a lawn chair on his porch typing on his laptop as photorealistic 4k digital art

Reading Time: 7 minutes

AI writing tools have left a tsunami of an impact across the academic and professional worlds over the last few months.

Tools like ChatGPT are mindblowing. You can ask a computer a question and have it make your essays, tables, or even code for you.

With this came a really important question – how do you check for AI writing?

To be completely transparent: You cannot officially prove something was written with AI. It's not like plagiarism. There's no actual watermark, you're really just looking at words on a page.

When you try to detect writing from ChatGPT, you'll essentially get a prediction score that tells you how well a robot would write the exact paragraph you're trying to scan. AI writing detection is fairly complex, but this explains it pretty well.

These machine learning algorithms & AI tools aren't capable of generating extremely high-level, creative writing yet. These detectors take advantage of that. If you submit writing that lacks creativity, has a complex vocabulary, and is full of patterns, there's a higher chance it was written with AI than not.

Of course, bad writing is still horrible. I've tested some of my own articles (and I promise I did not use ChatGPT), and they came back as AI-written. It's really murky water – and in a few months, it won't even be possible to predict anymore.

Anyone who tells you they can accurately detect AI-written content 100% of the time is lying. You can never be sure, it's not provable, as these are only predictions.

But sometimes, writing will get flagged as AI regardless of who wrote it. I've tried to outsmart these already-janky AI detectors and came up with a few tips to help humanize AI text to bypass AI detectors.

How to Bypass AI Writing Detectors

There are a bunch of things you can do, but the easiest is probably just to start with a word rewriter like Undetectable AI or HideMyAI (more words for free). Combine these with other syntactical changes, and you should be good to go. Here's a list of our best tips:

Use an AI Word Scrambler

Since AI models work on predictive text, you have to make text.... uhhh... unpredictable! I've had good success using tools like Undetectable AI to automatically reword my ChatGPT-generated writing into more human-looking content. Sometimes these tools actually just talk like a robot, no matter what you prompt it. Here's an example:

AI-produced text written with ChatGPT

So I asked ChatGPT to write me some human-looking stuff, and it kinda looks good. I threw the output into two AI detectors (Originality and Content at Scale) and both came back positive. I even used GPT-4 (which can write better than anything I've seen AI attempt before).

Came back 100% as AI. I threw this into Undetectable with the default settings and it came out with a result that was marked as human-written after my first attempt. Pretty awesome.

Definitely make sure to actually read your new paragraphs though. Depending on the complexity of the original content, you might notice some weird syntax in your new result.

I've put in some pretty scientific writing about a medical condition and the new response had some words in it (probably attempting to be synonyms) that made absolutely no sense!

If you aren't having luck with Undetectable, you could check out a few alternatives. I've also used QuillBot and they are really great for helping paraphrase content. QuillBot actually lets you rearrange the words in your paragraph (to swap out things like synonyms).

You can change the synonym level, language, and tone of voice with QuillBot as well. It's really a unique paraphrasing tool.

QuillBot AI Paraphrasing tool

Minimize Your Word Repetition

Another way of bypassing AI detection is to reduce the times you use the same word in an article. AI-generated content frequently repeats itself. Get rid of these words manually by proofreading and adjusting. This is basically part of what I mentioned above, but it's the human way of doing so!

Not all AI-generated content will repeat its words over and over, but the older models that have been trained on less parameters tend to write at lower levels. Lower levels = more predictable.

Attribute Your Sources and Verify Facts

AI often gathers content from the internet without discernment. Even worse, if it doesn't know what to talk about, it will just make things up.

Go ahead. Try to ask ChatGPT to write you a paragraph on the results of the 2026 Olympics. If it's in the mood, you'll see some crazy believable statistics. It's not true of course, but that's just how these tools are designed. It makes predictions based on what it knows. It'll look real, but it won't be.

So if you really want to differentiate your writing, cite your sources! And if you're using AI that has access to the internet (like the ChatGPT with browsing mode), you'll have references you can link within your articles to further humanize your writing.

Most of what AI is trained on isn't going to have the latest citations in its training data. This is some manual work you'll have to do but it makes you writing quality better & increases credibility. In-text citations are definitely the way to go

Make It Personal

By adding a few analogies or idioms in a piece of writing can almost certainly make it bypass any type of AI detection.

These detectors aren't THAT creative. They're wonderful and I won't discredit their potential generations, but again, they're still not real. Add a few sentences relating the topic at hand to a personal anecdote to increase your credibility and further help yourself bypass these detectors. Not to mention, it makes your writing a lot more relatable (isn't that the goal though!?)

Get Assistance from Another Human

So let's say you've done everything and your content is long.

You don't really want to rewrite everything, right?

Sometimes a great way of bypassing the AI writing detector traps is to have a friend take a look at your article. You don't need to tell them it was (or wasn't) written with AI, rather just asking for tips to make it more readable.

Another person can see discrepancies that you won't, especially if they don't have an ulterior motive because you just asked them to help make your content more readable.

Rewrite It Entirely By Hand

Ok, this might sound funky but hear me out – rewrite the content completely!

This might not make sense at first but we're still in a grey zone of what AI writing tools can actually do. Sometimes it's worth it to use for outlining our content, but not actually writing the complete thing.

If this sounds like the best use of your time (and you truly want to bypass AI detection) then have AI write your essay or report, then rewrite it as if you were speaking to a crowd (tailor this to the audience you'll have read your submissions).

Final Thoughts

It's an unfortunate situation that we even have to bring up bypassing AI detection tools. In a year, we won't have this issue ever again. It's just not going to make sense. Good writing is good writing.

Instead of trying to police the way people produce content, we should be focused on teaching ethical decision making. If you should be creating original content, do it. If your life efficiency can be improved with the help of a few AI writing tools, do it.

There's even been recent concern from parents as TurnItIn has released an AI detector without even informing teachers how it works. Is it even accurate? How do you claim something that's unprovable?

I'm almost sure we'll see the public perspective on AI writing shift from policing to embracing and promoting responsible usage.

These tools can not only save time but also foster creativity and innovation, and we're only at the start of the AI journey. Trust me.

In the future months, lets aim for a more collaborative environment, where human creativity and AI-powered assistance work hand in hand to produce outstanding content while adhering to ethical guidelines and maintaining the integrity of the creative process.

You can't replace human expertise, but you can totally optimize it!

Want to Learn Even More?

If you enjoyed this article, subscribe to our free newsletter where we share tips & tricks on how to use tech & AI to grow and optimize your business, career, and life.


Written by Justin Gluska

Justin is the founder of Gold Penguin, a business technology blog that helps people start, grow, and scale their business using AI. The world is changing and he believes it's best to make use of the new technology that is starting to change the world. If it can help you make more money or save you time, he'll write about it!

Subscribe
Notify of
guest

1 Comment
Most Voted
Newest Oldest
Inline Feedbacks
View all comments