- You will have noticed that this OpenAI technology is capable of camouflaging very well among humans
- The bot collects data from sources it was trained on and reorganizes it into simple sentences
- What are the tricks we have to detect the text written with ChatGPT?
If you have used Artificial Intelligence for some of your obligations, both student and work, you will have noticed that this technology is capable of camouflaging among humans in a very natural way. Also that it will improve over time and it will be increasingly difficult to separate from it. But how to apply to newsrooms? Supposing that Have you ever wondered if a text was written with ChatGPT and not by a human, in this article we show you how to find out quickly and easily so as not to be deceived by false content.
The fact is that although writing is an art, writing content is often a business. Since platforms such as blogs emerged to this day, there are a large number of writers trying to survive with their pen. Evidently this clashes with the capacity of Artificial Intelligence to impersonate human.
There are those who continue to try to attract us with words processed by their minds and written by their hands, but also those who have given in to the charms of AI to include its content in their texts.
How does ChatGPT create on-demand content?
To better understand what we’re talking about, we need to dig a little deeper into how ChatGPT creates content. Basically, this system is considered a revolution in the AI industry. But it is far from smart.
The bot collects data from sources it was trained on and reorganizes it into sentences that are not foreign to us. It takes advantage of its “knowledge” in grammar, and its results are usually better than those of other assistants. As a consequence, it is very difficult to tell at first glance whether it was written by a human or by ChatGPT.
So how to detect typed text with ChatGPT?
There are some tricks that could be very helpful if you are trying to find out the origin of a piece of content. Those who work as teachers or professors might be interested in collaboration tools specially created to detect typed text with ChatGPT. Let’s see what the instruments consist of.
Try the following methods to find out if the content was written by ChatGPT or a human:
Free or paid services
There are endless free Artificial Intelligence content detection tools available on the internet. They all have the same objective: to find out if the text was written by a user or if it was written by this technology.
Some paid tools – also those that offer a free trial version – can provide a specific score to know if the content was created with ChatGPT. That being said, false positives are not uncommon.
Paid content authenticity tools check for grammar and sentence formation. Now, if the writer uses ChatGPT to simply grammar check the text or translate it into a different language, the paid tools will give a false positive. For the above, you should not be 100% sure of their analysis.
Finally, you will have to follow the related news with new algorithms that have enough potential to detect content written by OpenAI. These mechanisms may improve along with ChatGPT itself.
Using ChatGPT on itself
ChatGPT does not create content, but adapts the information it has to a specific request from its user. It just rearranges it. Therefore, you can use this intelligence to prove the point itself.
All you have to do is follow a series of steps in which ChatGPT will show off its ability.
- Open ChatGPT and click New Chat to start a new conversation
- Ask him to write an article about the topic of that content
- If it is much shorter or longer than the text to be analyzed, ask for a specific extension
- Compare the two articles. If indeed the original is from ChatGPT, you will see that the information is rearranged
Of course, the more drafts you ask for on the subject, the more accurate your conclusions will be.
If whoever should have written that text usually has grammar errors, it is easy to discover the deception. There are many criticisms that can be made of ChatGPT, but if it does not fail at something, it is its impeccable grammar.
Just check the text using the Grammarly plugin. If the red flags can be ignored, you should assume that the text has already been passed through a grammar checker. Or, failing that, that the editor is very good grammatically or that he is born from the “pen” of Artificial Intelligence. Add the pieces of the puzzle.
Other frequently asked questions
Does Google accept content written by AI bots?
Although different sources give different answers to this question, it is better to avoid such content for a reason. When ChatGPT is asked to write an article on a specific topic multiple times, it will create similar content by rearranging the words. Google’s algorithm could detect it and mark that copied content. Eventually, if you have a blog or website, that could cause serious problems for your positioning.
Have you been able to detect text written with ChatGPT using these tricks? How was your experience?