Copy-pasting raw output from ChatGPT is the fastest way to lose your audience’s trust and attention. We have all reached a point of saturation where we can spot a machine-generated post from a mile away, usually within the first three words. I just saw this incredible post from an AI professional who has built a solution to this exact problem.
The creator of this tool recognized that the biggest issue isn’t the information AI provides, but the style in which it delivers it. Most people struggle to prompt the AI correctly to break its default habits. To solve this, the expert trained a specific, free custom GPT designed to strip away the robotic veneer and force the text to sound human. It is a brilliant workaround that focuses on unlearning the bad habits that Large Language Models (LLMs) are practically hardwired to display.
The Mechanism: Filtering Out the “GPT-Speak”
💡 The core philosophy behind the author’s tool is that AI is naturally risk-averse and polite, which leads to fluffy, empty writing. When you ask ChatGPT to write something, its default setting is to provide a comprehensive, neutral, and structurally repetitive answer. It tries to cover every base, which results in that distinct “AI flavor” we all dislike.
The original poster explains that this custom GPT works by enforcing a strict set of negative constraints. It isn’t just told what to write; it is explicitly told what not to write. By banning specific vocabulary, sentence structures, and hedging phrases, the tool forces the AI to get creative. It pushes the model to simulate human idiosyncrasies—like varying sentence length and taking a firm stance—rather than defaulting to the statistical average of all internet text.
📌 Insight 1: The Vocabulary and Filler Purge
One of the most valuable takeaways from the expert’s analysis is the absolute necessity of banning “obvious AI vocabulary.” You know the words I’m talking about. If a sentence contains “delve,” “tapestry,” “landscape,” “unlock,” or “unleash,” your reader immediately tunes out. These words appear frequently in AI writing because they are statistically common in the academic and marketing data the models were trained on.
But the author points out it goes deeper than just specific words; it is about “padding.” AI loves to fill space with empty phrases that add zero value. It will often start sentences with “It is important to note that…” or “In the realm of…” to bridge ideas. This is fluff. A human writer gets straight to the point. The creator’s method involves ruthlessly cutting this meta-commentary. You don’t need to announce you are making a list; you just make the list. You don’t need to summarize the previous paragraph; you just move to the next thought. By removing these buffers, the writing immediately feels more confident and less algorithmic.
📌 Insight 2: Breaking the Metronome Effect
Human speech is messy, varied, and rhythmic. AI text, by default, is a metronome. The LinkedIn user highlights that one of the biggest tells of AI writing is a lack of pacing variance. The AI tends to produce sentences of roughly equal length, one after another, creating a monotonous drone that puts readers to sleep. It feels perfect, grammatically, which is exactly why it feels wrong.
The expert suggests that to humanize text, you must actively disrupt this flow. You need a mix of very short sentences. Punchy ones. Followed by longer, more complex sentences that explain a concept in detail. Then back to short. This variation mimics the way we think and speak. When we are excited, we speak quickly (short sentences). When we are explaining, we slow down (long sentences). The author’s approach forces the AI to replicate this “jagged” rhythm, preventing the hypnotic, sleep-inducing cadence of standard outputs.
📌 Insight 3: Subjectivity and Concrete Details
Perhaps the most critical point the original poster makes is about “taking a stance.” AI is trained to be objective and helpful, which often translates to “hedging.” It will use phrases like “on the other hand,” “typically,” or “it depends,” trying to avoid being wrong. The problem is that human connection is built on shared opinions and concrete experiences, not neutral observations.
The creator emphasizes that you must stop hedging and start being concrete. Specific writing sounds human because humans live in a specific world; AI lives in a general one. Instead of saying “marketing strategies can be improved,” the expert advises you to say “this marketing strategy failed.” Furthermore, the author notes the importance of using “I” and “you.” Conversational writing creates a direct line between the writer and the reader. AI avoids the first person unless told otherwise. By injecting personal pronouns and concrete, specific examples rather than general abstractions, the text transforms from a lecture into a conversation.
The Nuance of Humanization
While this tool and these tips are powerful, there is a challenge to keep in mind. Over-correcting can sometimes make the text feel forced or overly colloquial in a way that doesn’t fit a professional context. The goal isn’t to make the AI sound like a teenager texting; it’s to make it sound like a professional talking to a colleague. You still need to review the output to ensure the “human” voice matches your specific human voice.
Checklist for Humanizing Your AI Text
Based on the brilliant list shared by this innovator, here is a quick checklist you can use to audit your own content before hitting publish:
- ✅ Kill the Filler: Delete “In conclusion,” “It is worth noting,” and “In today’s world.”
- ✅ Vary Length: Ensure you have a mix of 5-word sentences and 20-word sentences.
- ✅ Be Direct: Replace passive observations with active opinions.
- ✅ Audit Vocabulary: Ctrl+F for “delve,” “tapestry,” and “game-changer” and delete them.
- ✅ Use Pronouns: Rewrite passive voice to include “I” and “you.”
This resource is a massive help for anyone looking to maintain authenticity while using modern tools!
Check out the full post to access the free GPT and read the author’s complete guide.