Quick ChatGPT 'hack' instantly makes its outputs more 'human'

[Garth Gator]

One drawback about using large language models (LLMs) like ChatGPT is their responses get too robotic (well, what did we expect?). It turns out a nifty trick can nudge their responses towards a more natural, “human”-like conversation.

 

The secret lies in a simple fix to your prompt, as digital marketing expert Joe Benson suggests. He says instructing the chatbot to respond  in the tone of a text message to a friend works wonders.

 

 

Instead of a dry prompt like “Explain to me the meaning of life. Write a 40-50 word summary,” try this:

“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.”

 

This texting trick could be a boon for anyone who wants ChatGPT to generate emails, descriptions, and other types of content that sound like they came from a real person.

 

• https://bgr.com/tech/one-simple-chatgpt-hack-makes-ai-responses-sound-like-a-human-instead-of-a-robot/
• https://twitter.com/website_joe/status/1765597412859519106

 

 

Image: Rokas Tenys | Dreamstime.com