Students have already started using ChatGPT to assist them in writing essays like maniacs. Instead of spending the effort required to practice their writing skills and come up with their own ideas, they rely on ChatGPT for AI-written Essays. Teachers have been trying to find ways to counter this practice without much success.

In our previous article, we discussed 5 ways that ChatGPT could assist you with your work. In this article, we will show you what you need to do to find out if a student has cheated using AI software. 

Steps for identifying AI-written Essays

Up until now, the software we had at our disposal, even the commercial ones, such as Turnitin, could only help us discover plagiarised content, not AI-written ones. So, what do we do to counter AI-written essays?

Step 1. Look for issues with the content

The first thing you need to look for is signs of unnatural language, such as repetitive sentence structures or overly complex vocabulary, which are common characteristics of text generated by machine learning models.

ChatGPT can’t fully understand the context or meaning of the text it generates and, therefore, often repeats phrases or uses words that may not be suitable for the particular context.

Step 2. Compare the Essay with the student's previous performance

Next, you need to compare the student’s writing style and use of language throughout the academic year with the essay they have submitted, and you suspect trickery. For example, suppose a student has demonstrated a certain language and writing style throughout the academic year and suddenly gives you a significantly more advanced essay or an essay written in a completely different style from their usual one. In that case, this could be a sign that the student has cheated.

Still, this method can be subjective and prone to mistakes. So, what should you do next to confirm if there was cheating?

Step 3. Use an AI-detection software

Now, you have to fight back, using the same weapons as the potential perpetrator. These weapons are the AI-detection software and can increase the likelihood of identifying cheating.

Option 1: AI Content Detector

Let’s begin with our first choice, called AI Content Detector. To use it:

  1. First, go to https://writer.com/ai-content-detector/.
  2. Next, copy and paste the student’s essay within the “add some text” field.
  3. Then we click “Analyze Text”.
  4. After it’s done analysing, it will come up with a result labelled as AI detection score on the top right side of our screen. The AI Content Detector will point out how much AI-generated text was found within the essay.

Obviously, the AI Content Detector is not flawless. But it’s free and works OK, especially if the student hasn’t made any substantial edits to the generated Essay they received from ChatGPT. In any case, you can use AI Content Detector for free as long as the text you input is less than 1500 words. If you want more, you can always upgrade to a paid plan for 162$ per year.

Option 2: ZeroGPT

Our next choice is called ZeroGPT. ZeroGPT is free and was exclusively developed for identifying essays generated by ChatGPT.

ZeroGPT’s developers claim that it has a detection rate of around 98.5%.

Actually, I tested their claim. The results were incredible. ZeroGPT accurately identified all text written by ChatGPT’s AI during my test runs. It is even more potent than the AI Content Detector as it uses advanced algorithms to identify foul plays.

  1. To use it, first, you have to visit https://www.zerogpt.com/.
  2. Next, you copy and paste the text within the empty field. There’s no restriction to the number of words you can input. 
  3. Now you have to press the Detect Text button. ZeroGPT will analyse your input, and within a few seconds, it will come up with a result defining if the text was written either by ChatGPT or by a real person.

Bottom line

As the AI language models become more sophisticated, they make it increasingly hard to distinguish between human-written and AI-generated text. Thankfully, programmers have already started to come up with new software that can help teachers identify content that was not written by their students. Please note that the AI Content Detector and ZeroGPT aren’t the only choices available at the moment, but they are free and work well in most cases.

In conclusion, keeping an eye out for unnatural language and comparing a student’s writing style throughout the year, as well as using AI detection software, can help you identify AI-written content, at least on some level.