Photo by Franz26 on Pixabay

Introduction

To humanize ChatGPT content, you need to understand how AI language models generate content and find ways to break the patterns that lead to robotic writing.

AI language models are algorithms that are trained to follow specific linguistic patterns using predictive text technology. So they are not capable of independent decision-making or reasoning, and all they do is generate predetermined responses. That is why AI content sounds too stiff and repetitive.

While there is a way to make ChatGPT content sound more human, it still won’t get close to actual human creativity and display of emotions.

With that said, this post will aim to help you generate MORE human-like content from AI, but you should edit it with your own writing if you want actual human content.

Why Do We Need Human-Like Content?

In many cases, ChatGPT can generate clear and concise content better than humans. For example, if you are writing an email or translating a document, ChatGPT would do it faster and more accurately. And it wouldn't matter if it's AI-generated. But there are cases where it would, like writing a blog post, for example. If you want your blog to be unique and rank well, you have to focus on original content that sounds human. That’s mainly because your visitors would connect better with real, authentic experiences and writing than a bot.

Photo by Google DeepMind on Pexels

How to Break AI Patterns

There are three main areas of focus when it comes to prompting ChatGPT: perplexity, burstiness, and syntax.

Perplexity

Perplexity is a measure of how surprised or confused you might get when reading a sentence. Imagine you have to guess the following sentence: “The cat is chasing the…” If you say “mouse“, it’s not very surprising, so that’s low perplexity. However, if the answer is “lion,” that’s very unexpected and hard to guess, which means it has high perplexity. Humans generally write with higher perplexity compared to ChatGPT because AI lacks human depth and creativity.

Burstiness:

Burstiness is the variation in complexity, structure, length, and pacing of sentences. Human writing varies more from sentence to sentence, creating a less uniform and repetitive style. For example:

Low Burstiness: "The cat chased the mouse. The mouse ran away. The cat got tired and slept."

High Burstiness: "The cat dashed after the mouse, leaping over obstacles, skidding around corners, and causing havoc in the kitchen. The mouse darted frantically, zigzagging to evade capture. Eventually, the cat, exhausted by the wild chase, collapsed onto the floor, panting, and promptly fell asleep."

Unlike AI, the human mind isn't limited by programming data and can write with unlimited creativity. AI-generated content has improved but still follows recognizable patterns.

Syntax

Syntax involves arranging words correctly to make sentences clear and easy to understand. For example, “The dog runs quickly." is a correct syntax. But "Quickly runs dog the." isn't.

Human writing often breaks traditional syntax rules based on the context and the purpose of writing. For example, starting sentences with “And“ or “But“ for a conversational tone or adding emphasis like, “And then—BOOM!—the fireworks exploded, lighting up the night sky."

ChatGPT is more consistent in following grammar and sentence structure rules, which results in repetitive and stiff content.

How to Prompt ChatGPT to Write More Human

Now that you have an idea of how AI text generation works and the difference between AI and human writing styles, you can direct ChatGPT to create unpredictable, varied, and freestyle content using simple prompts. Or you can provide a human text as a sample and ask ChatGPT to mimic the writing style. But keep in mind that short prompts are more effective and don’t overwhelm ChatGPT with excessive information.

I have spent a fair amount of time experimenting with different prompts until I came up with the one that rules them all. It’s short, to the point and instructs ChatGPT to maximize its effort in breaking patterns.

Re-write in 100% human perplexity, supercharged burstiness and unpredictable syntax

Let me show you how effective this short prompt is with a real example from AI-generated texts.

Original AI text:

“ChatGPT advances in mimicking human writing through improvements in natural language processing, vast data training, and fine-tuning for nuance and context. It studies patterns, word choices, and sentence flows across a variety of texts to craft responses that feel more relatable and authentic. Techniques like transformer-based architectures allow for better understanding of subtle cues, emotions, and conversational dynamics. The system evolves with feedback, adapting to user preferences and creating a balance between coherence, unpredictability, and a human-like tone for more realistic and engaging outputs.”

After using the prompt, “Re-write in 100% human perplexity, supercharged burstiness, and unpredictable syntax,”:

“Wow, ChatGPT is totally rocking the human writing game! It's seriously amping up its skills in natural language processing, soaking up loads of data, and getting all fine-tuned to nail those little details and context. It's diving deep into patterns, picking out just the right words, and dancing through sentences in all sorts of texts to give us responses that really hit home and feel totally genuine. With these cool transformer-based architectures in the mix, it's even getting a better grasp on those subtle cues, emotions, and the way conversations ebb and flow. And get this, it's constantly evolving based on feedback, adapting to what users like, and finding that sweet spot where it's coherent but also keeps us on our toes with its unpredictable and totally human-like vibe. You won't believe the realistic and engaging stuff this baby churns out!”

Use AI Humanizer Tools

There are many AI tools in the market that do a decent job at humanizing ChatGPT. But because that is a costly solution (subscriptions can go up to $45 per month), I wouldn’t use them unless I really need my content to bypass AI detectors (which in most cases isn’t ethical).

Check out Katteb; it has a pretty decent AI humanizer. Katteb also features a "terms" option that lets you specify which words or phrases to eliminate and which alternatives to use. This gives you more control and allows you to write in your own style.

Go ahead and experiment more. Share your prompt in the comments and let us know how it worked for you!