ChatGPT, like all good LLMs (Large Language Models), is particularly good at writing essays. OpenAI’s advanced AI language model can generate human-like text from written prompts. Its capabilities extend to everything from helping people communicate more effectively, to assisting with customer service, and even as a writing tool. Yes, that includes your homework. However, AI checkers may be able to detect your use of OpenAI’s AI chatbot, including the commonly used plagiarism prevention system, Turnitin. In this article, we’ll look at how Turnitin can be used to detect AI and what that means for students planning on using the service.

✓

Quick Answer

Paired with its plagiarism checker, Turitin’s AI writing indicator can detect all forms of written work that have been assissted by AI tools like ChatGPT.

Turnitin can detect ChatGPT

Students across the world have been taking notice of ChatGPT’s abilities. There is the temptation for many students to use the chatbot (among other AI writing tools) to write their essays. At the same time, the need for an AI detector has sprung up among educators to combat this practice. Academic integrity aside, there are some risks with using ChatGPT on work you plan on submitting.

Yes, Turnitin is one such AI detector.

Turnitin has a new AI writing indicator that can detect AI-generated content with up to 98% accuracy. Launched on April 4th, 2023, it even has the detection capability to flag AI-assisted content. The AI tools join Turnitin’s plagiarism checker to ensure that, even as technology advances, LLMs aren’t the reason our future doctors and engineers ‘faked it till they made it’.

Annie Chechitelli, Chief Product Officer at Turnitin, is confident in the AI-checking capabilities of the software.

“Our model has been trained specifically on academic writing sourced from a comprehensive database, as opposed to solely publicly available content. As a result, Turnitin is more tuned to finding instances of potential dishonesty in student assignments.”

Annie Chechitelli, Chief Product Officer at Turnitin

How does Turnitin work?

Turnitin is a plagiarism detection software application that educational institutions widely use to check for plagiarism in student assignments. It compares the work to a comprehensive database of previously submitted work and runs an originality check.

Moreover, it compares that work to other sources on the internet to identify any instances of plagiarism. Turnitin has become a valuable tool for instructors and professors to ensure academic integrity. While plagiarism detection by itself is not an AI-powered process, the software’s new AI writing indicator adds a new tool to the educator’s arsenal.

The detector is currently only available to non-students and educational institutions. And, only instructors and administrators are able to see the indicator score in the report.

Using Turnitin to detect AI

AI detectors check for AI-generated text with several different methods. In addition to the more subjective methodologies mentioned in this article, some companies are working on an “AI signature”. An AI signature will be an invisible addition to text generated by an AI chatbot that other AI detection systems will be able to check for to conclusively prove the use of AI.

Turnitin can detect ChatGPT by analyzing individual sentences and running them through its database, ultimately providing an AI detection score that ranges from 0% to 100%.

Achieving 100% accuracy is exceedingly difficult, largely due to false positives. AI detection software can falsely claim to detect AI-written text that was in fact human-written. This issue extends to all AI detection software. It is not exclusively a Turnitin AI detection problem. The problem is somewhat philosophical – after all, how do you tell the process of creation apart by the end result alone, when the end result is quantifiably identical?

Essential AI Tools

Has anyone been caught using ChatGPT?

Yes, students are often caught using ChatGPT. Reported cases have come to light where professors have recognized the use of ChatGPT, Either through tools like Turnitin or through the style of writing, which might deviate from a student’s usual style. Educators may also notice the presence of exceptionally well-crafted points that lack cohesion between themselves.

In addition, there are often false positives – teachers ‘catching’ students that did nothing wrong. This is because AI checkers are just as imperfect as the AI students use to cheat.

Being caught by markers when handing in work that has been written by AI is definitely a factor that should be considered when using ChatGPT. With more people using AI as a helpful tool for writing, there have been more and more instances of students being caught. It’s important to weigh up the risks of using AI before handing in anything that could be detrimental to your future education.

Can teachers detect AI?

Other than AI detection software, there are methods that instructors can use to identify the use of artificial intelligence in student assignments. One method is to analyze the language and style of the writing. ChatGPT has a particular style and may use certain phrases or words that are not typical of a student's work.

Instructors can ask students to explain certain concepts or ideas in their own words. In doing so, they can determine if the students truly understand the material.

In addition to this, students are expected to submit their work in a given format. The APA (American Psychological Association) provides guidelines for making text easy to read, and especially making essays neat and consistent for educators who will need to read many documents in a short space of time. If your AI chatbot can’t write to APA guidelines, that’s a huge red flag.

Conclusion

So, can ChatGPT be detected by Turnitin? Well, yes it can! While this may concern some, students need to understand that using artificial intelligence in this way is not a replacement for their academic work and critical thinking skills.

It is important for instructors to be aware of the potential use of ChatGPT and to use detection systems to overcome it. Ultimately, students must consider the ethical implications of using ChatGPT, and so do the instructors.

To find out more on this topic, we’ve covered the use of other ChatGPT detectors here.