There are plenty of supposed ChatGPT “tricks” floating around online, and I tend to approach most of them with a healthy dose of skepticism. If you've used ChatGPT for any length of time, you know that it’s getting better at sounding less robotic, but it still struggles with the overuse of words and phrases like, “that distinction matters,” “increasingly” and the overuse of patterns like “if x than y.”
But for those who spend a lot of time writing emails or corporate notes and use ChatGPT has a faster way to get the job done, than you might be interested in this viral trick. By simply adding /human before your prompt, you can get ChatGPT to ditch its telltale AI writing habits and respond more like a real person might.
To understand the reasoning behind this trick, it’s important to know how large language models respond to context. A single instruction like /human can sometimes steer an answer in a way that you’d expect and less like AI.
To test this trick, I gave ChatGPT five everyday writing tasks, from drafting an awkward email to writing a LinkedIn post that's practically begging for AI clichés. Here’s what happened.
1. The awkward work email
Prompt: Write an email telling my boss I won't finish a project by Friday and need until Monday. I don't want to make excuses, but I also don't want to sound careless.
After running the identical prompt with /human, the biggest difference is that it removed what sounds like a carefully engineered “professional email” and more like a person giving their boss an early, straightforward heads-up. It actually sounded like something a real human might send.
2. The birthday text
Prompt: Write a birthday text for my best friend of 20 years. Make it warm and funny without getting overly sentimental.
You’ll notice the /human version changed mostly the cadence and specificity, not the message itself. It made the message messy in a deliberate way. And, it uses more conversational rhythm: short sentence > joke > longer thought > joke > affection > done. It’s not overly polished in the way ChatGPT usually responds. There are also far less generic sentiments.
3. The recommendation
Prompt: Recommend a coffee maker for someone who drinks two cups every morning, hates complicated appliances and has a budget of $150. Explain your recommendation in about 150 words.
Adding /human made the text sound more like a person answering the question from experience. The original sounds like a product recommendation from the actual company, which makes it sound less sincere. The human version also made fewer transitions. It removed obvious AI phrases like, “more importantly” and “for your needs,” which are typical AI fluff.
4. The personal story
Prompt: Write 150 words about the moment you realize your oldest child is growing up faster than you expected. Make it emotional without being overly sentimental.
The /human version feels less polished and more conversational, replacing carefully constructed imagery and symmetrical phrasing with messier, more specific observations. It also avoids neatly explaining the emotion, letting humor, uncertainty and slightly imperfect phrasing make the moment feel more natural. It shows rather than tells, which is what human writing often does.
5. The LinkedIn post
Prompt: Write a LinkedIn post announcing that I'm starting a new job. I'm excited about the opportunity, but I don't want it to sound like every other LinkedIn announcement.
The /human version drops the LinkedIn cliches and replaces them with more specific, slightly messy thoughts like nerves, self-doubt and worrying about a first-day outfit. The sentence structure is also looser and more conversational, so it reads more like someone sharing news than crafting a professional announcement. This test definitely felt like the final boss of AI writing.
Final thoughts
It’s clear that this viral trick definitely works, even with the latest version of ChatGPT, which already claims to cut the fluff. But even by adding /human to the prompt, AI writing is still AI writing. No matter how much you try to adjust a draft written by AI, it’s always better to start with a human draft and then ask AI to help you edit. Once you start with an AI draft, it’s very hard to compose it into something that doesn’t sound completely robotic.
Give this trick a try and let me know what you think in the comments. Have you tried another viral hack? Share your thoughts and start the conversation.
Follow Tom's Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds. Subscribe to Tom's Guide on YouTube and follow us on TikTok.
More from Tom's Guide
- I asked ChatGPT, Gemini, Claude and Muse to predict how bad the Nor'easter is going to be this weekend — here's what they say
- ChatGPT-6 or Claude Opus 5.5? Here’s which AI I'd use for writing, planning and 5 other everyday tasks
- I tested ChatGPT-6 vs Claude Opus 5.5 with 5 everyday prompts — it wasn’t even close
