Artificial Intelligence

Five Telltale Signs That a Blog is AI Generated

Our Roving Reporter

Last month our roving reporter published an article where Chat GPT was pitted against a human writer. Following up on reader feedback here are some five telltale signs, that the content may have been AI-generated. This is not an exhaustive list, and this article is written in an Australian context.

1. Check the spelling and grammar. Chat GPT is trained to write in almost flawless US English. So look closely has it spelt organisation or organization? Colour or color? Also, check for the use of the Oxford comma. It is less prevalent in US English than in Australian or even UK English.

 

2. Repetition of words and phrases. On the surface AI wants to present itself as the expert, however, it lacks the knowledge, of a real human expert with years of experience has, in that particular field. Generative AI has the tendency to write filler paragraphs that repeats the points made previously in the text.

 

3. Inaccurate information. This one, most people already know. Generative AI tools like ChatGPT, are only trained on information up to November 2021. Any news or latest trends from 2022 or 2023 will see it struggle.



4. Lack of colloquial language. Similar to point one, generative AI is trained to write in a formal tone. Any use of humour, nuances or casual language is absent. So have a closer look does it write uni or university? Cop or police officer? 



5. Lack of citations. Anyone who writes for a living knows that you must fact-check and acknowledge your original sources. Unless prompted Chat GPT and other generative AI tools will almost never cite where they got the information from. 

 

And there you have it. These are some of the signs to look out for. Is there anything else that we have missed? Please let us know. Just a disclaimer, a human wrote this blog.