Chatting with ChatGPT is part of my daily life these days. I talk to ChatGPT for research, I do it when I have to “google” something, and sometimes I do it for the lols. Since I’m on ChatGPT Plus, I also get to experience OpenAI’s best generative AI features as soon as they’re available in Europe, where I live. 

However, not all improvements come from OpenAI directly. Issuing the right prompt is key to getting the most out of ChatGPT. And I’m always ready to test interesting ChatGPT hacks that can get me better replies. The newest one concerns the tone of ChatGPT’s responses. Specifically, you can get the chatbot to sound more human than robotic with a simple trick.

I don’t think I mind ChatGPT’s tone. It’s something I got used to. Maybe the tone being a little robotic is something I need to ground me. Yes, chatting with AI is as easy as talking to a human. But ChatGPT’s tone reminds you this is the AI you’re talking to, not your buddy.

Still, the following ChatGPT hack, courtesy of Joe Benson on X, involves a simple addition to your prompts.

The ChatGPT tone hack

If you want ChatGPT to write some content for you that sounds natural instead of machine-generated, all you need to do is add one simple instruction to the end of your prompt:

[prompt] in the tone of a text message to a friend without a salutation or emojis

Benson offered the following prompt example: 

Regular Prompt:

Explain to me the meaning of life. Write a 40-50 word summary.


The meaning of life is a philosophical question that has varied interpretations. Some view it as the pursuit of happiness and fulfillment, others see it as contributing to the greater good, and many religious and spiritual beliefs tie it to a higher purpose or divine plan. Ultimately, it’s a deeply personal question, and its answer varies widely depending on individual beliefs, values, and experiences.

Then, he added the “secret sauce”: 

Prompt with Secret Sauce:

Explain to me the meaning of life. Write a 40-50 word summary in the tone of a text message to a friend without a salutation or emojis.


The meaning of life? Big question! It’s really what you make of it. Some find it in happiness, others in making a difference. It’s all about your personal journey and what makes you feel fulfilled. Think of it as your own unique adventure.

The second example certainly reads better and sounds more personal than the initial example. Will it work for everything you ask ChatGPT to do for you? Well, that’s up to you to try. 

I went to the chatbot with a different challenge using the trick above. I asked it to give me the result of a simple math problem. I must say the tone hack version sounds better: 

I used the ChatGPT tone hack to solve this math problem.
I used the ChatGPT tone hack to solve this math problem. Image source: Chris Smith, BGR

You can make the hack even better

If this ChatGPT hack works for you, there’s another trick that can make things even easier. Rather than adding “in the tone of a text message to a friend without a salutation or emojis” to every prompt you write, you could add that line to ChatGPT’s custom instructions. That way, it’ll keep the tone for all of its responses. 

Or maybe you want to use the new ChatGPT memory feature to have the chatbot remember the tone you want for prompt replies.

Also, you don’t have to use the command above. You can tweak the “personality” of ChatGPT with any instructions. As long as the AI gets the job done, the tone will not really matter. ChatGPT could be Captain America for all I care: 

What if ChatGPT were Captain America?
What if ChatGPT were Captain America? Image source: Chris Smith, BGR

That said, I’ll have ChatGPT keep its default tone for my chats. But it’s up to you to decide what ChatGPT should sound like.

[ For more curated tech news, check out the main news page here]

The post One simple ChatGPT hack makes AI responses sound like a human instead of a robot first appeared on

New reasons to get excited everyday.

Get the latest tech news delivered right in your mailbox

You may also like

Notify of
Inline Feedbacks
View all comments

More in computing