How to Change AI Text to Human Text in 5 Steps

At Editly AI, we have seen that, with the recent rise in AI-generated text, has come a concomitant wave of interest in its humanization. Educators and students have waged a virtual war armed with AI text generators and AI text detectors. Professionals and academics have weighed in with ethical concerns, while techies have tried to make a quick buck with apps that “humanize” AI text by replacing random words with questionable synonyms. 

Yet, there seems to be very little practical advice out there for those that are looking to reap the benefits of AI-generated text by making it their own—or, in other words, making it human. While leaving aside the debate on what exactly makes text human, this post rests on the idea that a large part of what marks AI text as AI is uniformity and repetition. From this starting point, it offers you 5 practical and actionable steps to change AI Text to (more) Human Text. To do so, it examines a ChatGPT-generated text response to the question: “How do I change AI text to human text?”.

Step 1: Find Repetition in the AI Text

At this stage, your mission is to find the repetition that seems to plague AI-generated texts. More specifically, your focus should remain fixed on nouns (people, places, and things), verbs (actions), and adjectives (descriptors). Here, we find the repetitions in the sample AI-generated text (in bold):

Converting AI-generated text to appear more human-like involves several nuanced steps to infuse the text with elements of human writing styles, personal nuances, and the imperfections that characterize everyday communication. Firstly, introduce a more conversational tone, using colloquial language and contractions to mimic natural speech patterns. This makes the text feel more relatable and less formal. Secondly, incorporate personal anecdotes or hypothetical examples that add depth and relatability, as human stories often include personal experiences or emotions. Thirdly, include rhetorical questions, which are commonly used in human communication to engage readers and encourage them to think more deeply about the topic. Additionally, it’s beneficial to vary sentence structure and length, creating a rhythm more reflective of spoken language, where pacing can vary widely. Intentionally integrating minor imperfections, such as the occasional use of filler words like “um” or “well,” can also make the text seem more human, as it mirrors natural imperfections in human speech. Lastly, consider the use of humor or wit, which adds personality and warmth, making the text feel as if it’s coming from a human rather than a machine. By implementing these strategies, AI-generated content can be transformed to more closely resemble the nuanced and imperfect nature of human writing.

As you can see, we are not attempting any high-level matching. This much repetition is very characteristic of AI text, but not so much human text.

Step 2: Find the “To Be” Verbs in the AI-Generated Text

By “to be” verbs, we mean quite simply “am,” “is,” “are,” “was,” “were,” “be,” and “being.” The reason that we target these verbs involves not only repetition, but also uniformity of syntax. In fact, the overuse of “to be” verbs very often leads to these undesirable AI text characteristics. Below, we identify the “to be” verbs in our AI-generated text (in italics and bold):

Converting AI-generated text to appear more human-like involves several nuanced steps to infuse the text with elements of human writing styles, personal nuances, and the imperfections that characterize everyday communication. Firstly, introduce a more conversational tone, using colloquial language and contractions to mimic natural speech patterns. This makes the text feel more relatable and less formal. Secondly, incorporate personal anecdotes or hypothetical examples that add depth and relatability, as human stories often include personal experiences or emotions. Thirdly, include rhetorical questions, which are commonly used in human communication to engage readers and encourage them to think more deeply about the topic. Additionally, it’s beneficial to vary sentence structure and length, creating a rhythm more reflective of spoken language, where pacing can vary widely. Intentionally integrating minor imperfections, such as the occasional use of filler words like “um” or “well,” can also make the text seem more human, as it mirrors natural imperfections in human speech. Lastly, consider the use of humor or wit, which adds personality and warmth, making the text feel as if it’s coming from a human rather than a machine. By implementing these strategies, AI-generated content can be transformed to more closely resemble the nuanced and imperfect nature of human writing.

Please note the “to be” verbs hiding inside the contractions: “it’s” = it is. Now, we are ready to move on to the next step.

Step 3: Find the Transitions in the AI Text

By transitions, we mean the connectors that join sentences together. While too many exist to list here, they are often located at the beginning of sentences and followed by a comma. AI-generated text often relies too much on repetitive or cliched transitions. Here, we find the transitions in our AI text (underlined):

Converting AI-generated text to appear more human-like involves several nuanced steps to infuse the text with elements of human writing styles, personal nuances, and the imperfections that characterize everyday communication. Firstly, introduce a more conversational tone, using colloquial language and contractions to mimic natural speech patterns. This makes the text feel more relatable and less formal. Secondly, incorporate personal anecdotes or hypothetical examples that add depth and relatability, as human stories often include personal experiences or emotions. Thirdly, include rhetorical questions, which are commonly used in human communication to engage readers and encourage them to think more deeply about the topic. Additionally, it’s beneficial to vary sentence structure and length, creating a rhythm more reflective of spoken language, where pacing can vary widely. Intentionally integrating minor imperfections, such as the occasional use of filler words like “um” or “well,” can also make the text seem more human, as it mirrors natural imperfections in human speech. Lastly, consider the use of humor or wit, which adds personality and warmth, making the text feel as if it’s coming from a human rather than a machine. By implementing these strategies, AI-generated content can be transformed to more closely resemble the nuanced and imperfect nature of human writing.

Now that we have identified the transitions in the AI-generated text, we can proceed to the corrective steps to change our AI text to a human text.

Step 4: Eliminate Repetition and “To Be” Verbs in the AI-generated Text

When we speak of eliminating repetition in AI text, we are using density as our guide. In other words, the closer together that we find repeated words, the greater the urgency to get rid of the repetition. Conversely, the further away the repetition, the less the need to eliminate the repetition. 

In regard to the “to be” verbs, our mission is much more clear-cut. We attempt to eliminate as many of them as possible. For, they foster repetition in syntax and diction (word choice), as well as passive voice.

Step 5: Eliminate or Replace Transitions in the AI Text

Due to the great variety of transitions and their utilization contexts, this step is generally less straightforward than the previous steps to humanize ai text. Given our AI text, however, we can safely state that numerical transitions give a rather more robotic impression. Therefore, they are ideal candidates for replacement.

Result of Changing the AI Text into Human Text

Before we unveil our human rewrite of the AI-generated text, let’s revisit the original AI text for the sake of comparison. Here, we have combined repetition, “to be” verbs, and transitions.

Converting AI-generated text to appear more human-like involves several nuanced steps to infuse the text with elements of human writing styles, personal nuances, and the imperfections that characterize everyday communication. Firstly, introduce a more conversational tone, using colloquial language and contractions to mimic natural speech patterns. This makes the text feel more relatable and less formal. Secondly, incorporate personal anecdotes or hypothetical examples that add depth and relatability, as human stories often include personal experiences or emotions. Thirdly, include rhetorical questions, which are commonly used in human communication to engage readers and encourage them to think more deeply about the topic. Additionally, it’s beneficial to vary sentence structure and length, creating a rhythm more reflective of spoken language, where pacing can vary widely. Intentionally integrating minor imperfections, such as the occasional use of filler words like “um” or “well,” can also make the text seem more human, as it mirrors natural imperfections in human speech. Lastly, consider the use of humor or wit, which adds personality and warmth, making the text feel as if it’s coming from a human rather than a machine. By implementing these strategies, AI-generated content can be transformed to more closely resemble the nuanced and imperfect nature of human writing.

The original AI text weighs in at a total length of 215 words. Let’s now take a look at our humanized AI text from our Editly editors:

Humanization of AI-generated text involves infusing a text with elements of human writing, personal nuance, and common communication imperfections. To begin, introduce a more conversational and relatable tone that utilizes colloquial language and contractions to mimic natural speech. Then, incorporate anecdotes or hypotheticals to add depth through emotions and experiences. To further engage readers and encourage deeper thought about the topic, include rhetorical questions. In addition, vary sentence structure and length to create a mixed rhythm more reflective of spoken language. Intentionally integrating minor imperfections, such as the occasional use of filler words like “um” or “well,” may also make the text more human. Finally, consider using humor or wit to bolster personality and warmth. By implementing these strategies, you can successfully transform AI texts into human writing.

By following the 5 simple steps to changing AI text into Human text, we have reduced the total word count from 215 words to 128 words—a reduction of more than 37% in word count, without losing any of the original ideas. This math provides insight into the effectiveness of our steps in reducing the characteristic repetition and verboseness of AI text. Beyond what the math can capture, however, lies a variety and vitality that is simply not present in the AI text.

The Takeaway

You, too, can humanize your AI texts by following the same 5 steps. But, if you would like an experienced helping hand, then don’t wait to get started on Editly.