An email, an assignment, a cover letter… so many things can now be written with artificial intelligence, especially with ChatGPT.
But how can you tell if a text was written with AI? First, we can recommend our online ChatGPT detector, which we offer for free on Yiaho. But if you don’t have a computer or phone available, it’s still easy to spot language patterns that are used a lot by ChatGPT.
In this article, we’ll look at words, sentence structures, and expressions that are very often used by text generators.
Spotting a text written with ChatGPT: 4 foolproof tips
You can detect ChatGPT in a text by looking in the right places. Here are 4 signs that should raise a red flag!
1: The words used
ChatGPT tends to overuse expressions like “moreover”, “therefore”, and other logical connectors—handy for structuring an argument, but heavy to read! These expressions, while functional, are often too frequent in AI-generated writing.
Be especially alert for the phrase “moreover”, which shows up frequently and can be a good indicator that ChatGPT was used!
2: Stock phrases
Among the most common are stock phrases that feel a bit heavy. Here are a few examples:
- It is crucial to understand,
- It is important to note,
- Diving into the details,
These turns of phrase reflect a standardized style that can give away the text’s origin. ChatGPT loves starting sentences like this.
3: The structure of the text
A text written by ChatGPT can often start with a very basic introduction, such as: “In a world where (topic of the text) is constantly evolving.” Or “In the fascinating world of (topic of the text).” In general, ChatGPT introductions are quite school-like, lack originality, and make the writer sound like a total newbie!
Even worse: the conclusion. If a text ends with a paragraph simply titled “Conclusion” that summarizes almost word for word what was said before, without offering any new angles to think about, it’s very likely the text was written with ChatGPT.
4: A cold, school-like, very basic tone
Let’s be honest: asking ChatGPT to reply to an email or do an assignment can produce decent results. There won’t be any mistakes, and everything will be structured.
But it will seriously lack humanity, warmth, and personality! A text written with ChatGPT is generally cold and impersonal. And if the writer asks ChatGPT to write in a cheerful or “cool” way, the result can end up looking a bit ridiculous.
So it’s crucial to learn to spot the “robotic” and overly formal side of AI-generated text.
Also read: Chat GPT Detector: Here’s the most effective tool
Our plagiarism detector, just to be sure!
You now have four practical tips and concrete advice for detecting a text written with AI. But at Yiaho, we also have tools to give you extra certainty!
Your first reflex, once you’ve read the text and you’re unsure, is to use our detection tool. You’ll know exactly which part of the text was written with ChatGPT!
Finally, another simple tip: if, before ChatGPT came along, your colleague or student struggled to write an email or an essay, and since 2022 their writing has been flawless, that’s definitely something to wonder about!
Keeping these points in mind, you’ll be better equipped to recognize a text generated by ChatGPT. And you—have you ever caught someone openly using ChatGPT? Share your thoughts in the comments or on The AI Forum!
