ChatGPT or human text? Here's one surprising way to tell the difference

Summary

This article by Sabrina Ortiz, writing for ZDNet, makes the following points:

• Cheating has created a demand for AI-generated text detectors.
• Two experiments prompted ChatGPT to generate restaurant reviews.
• The study showed that ChatGPT describes experiences rather than sharing feelings and avoids personal pronouns.
• It also never uses aggressive or rude language and is programmed to avoid using irony, sarcasm or metaphor.

References

Recent Runway Posts related to this topic:

• Articles about AI-based ChatBots | The Runway (airforce.gov.au)
• Alternative Link: Articles about AI-based ChatBots | The Runway (airforce.gov.au)
• See all Runway Collections: Collections | The Runway (airforce.gov.au)

References from the Web:

• DEC 2022 How to spot AI-generated text—MIT Technology Review
• FEB 2023 ChatGPT maker OpenAI has a free tool that can spot AI-written text—ZDNet 

Source Information:

• Article Source: https://www.zdnet.com/
• Media Check: ZDNet - Media Bias Fact Check (mediabiasfactcheck.com)
• RAAF RUNWAY: RATIONALE, GUIDELINES, LEARNING OUTCOMES, ETC.