ChatGPT is the perfect example of how technology can be a double-edged sword. When Open AI launched ChatGPT in November 2022, it was received with a lot of enthusiasm due to its many capabilities. Even though this AI chatbot has many wonderful uses, some students are using it as AI writing tool to write essays and other academic works. This is a huge blow for universities and colleges because it’s being portrayed as an act of cheating.
The good news, however, is that recently, Turnitin introduced an AI content detector. This software claims that it can identify if an essay or article was produced through ChatGPT. Does it really work? That’s what we are going to find out in today’s article.
Before we answer the question – does Turnitin detect ChatGPT? Allow us to introduce this software. After all, not everyone is familiar with academic tools.
Turnitin is a web-based similarity index checker that looks for matches between a student’s assignment and other published works. Turnitin was developed to curb cheating in universities where some students would copy other people’s writings and submit them for marking.
Apart from detecting plagiarism, this tool also helps students avoid such incidents so that they can improve their academic writing. Turnitin usually generates a report of the assignment which is made available to the lecturer or member of staff. And it has access to a massive database of journals, articles, and catalogs. That’s why it’s so effective at identifying matching content.
Educators can now take a sigh of relief because Turnitin is capable of detecting ChatGPT-generated content with an accuracy of 98%. There have been arguments that essays written by ChatGPT can’t be flagged for plagiarism because they are not published anywhere on the internet. However, Turnitin has been programmed to identify different forms of AI writing. Even though the management behind Turnitin is yet to explain how they can do this. We are aware that it’s a project they have been running for almost two years and it’s now available.
From our analysis, Turnitin is capable of detecting ChatGPT content because both tools have access to a large database of content. Therefore, essays written using the AI chatbot will be drawn from similar sources.
Even though Turnitin can detect ChatGPT-generated essays and papers, you should know that AI technology is continuously being advanced. With time, you can expect ChatGPT to use more natural and complex language which won’t be easily detectable. Hopefully, by then, Turnitin will have also advanced its AI-detection software.
As mentioned earlier, Turnitin has declined to disclose how the software works in detecting AI-assisted writing. This is obviously due to competitive purposes because there are other AI detectors out there that may imitate their system. However, what we do know is that Turnitin also uses AI which scans assignments and compares them to a database of online data.
The Turnitin software identifies unique AI-generated patterns which are common with chatbots like ChatGPT. Even though ChatGPT tries to emulate human conversation, in some ways, it’s a bit different from human writing and that’s what Turnitin flags as AI content.
This is how Turnitin works when detecting AI content. When you submit a paper, it’s broken into segments of text that are divided into a couple hundred words. Each segment is run through the AI detector and given a score of 0 if it’s generated by humans and 1 if it appears to be ChatGPT created. A total score is given at the end depending on how many sentences appeared to be written by ChatGPT.
On their official website, Turnitin highlights that they strive to treat students fairly with new AI detection tool. And that’s why they only identify texts that seem to be AI-generated. Also, the reports for ChatGPT content are only available to instructors and account admins.
As a professional in the education sector, you need to regulate the use of ChatGPT by your students. Though useful, this AI chatbot can significantly impact a student’s analysis and critical thinking skills. In many institutions, the use of ChatGPT is a form of academic dishonesty that needs to be managed before it spirals out of control. Here are some of the ways experts can detect the use of ChatGPT;
🕵️ Use AI-Detector Tools Like Turnitin
So far, hundreds of colleges and universities have adopted the use of AI detection tools like Turnitin to check for content created using ChatGPT. Grammarly also has an AI detector. Running papers through these tools is an easy and quick way of checking for original work.
🕵️ Look for Similar Points of View from Students
Another smart way of identifying the use of ChatGPT is by looking for similar or generic views from students. If several of your students used ChatGPT, there is a high chance that their essays will look similar or share common sub-topics.
🕵️ Check the Writing Style
As a natural language processing model, ChatGPT may attempt to the best of its abilities to mimic human writing. However, if you look at its content keenly enough, you may notice some unique patterns. If your students’ papers contain complex words and sentences and they are a bit younger, they may have used ChatGPT.
🕵️ Ask More Questions about Their Work
If a student used ChatGPT to generate an essay or paper, they will probably have a limited understanding of the content. Therefore, they won’t be able to provide correct answers to questions related to their work.
🕵️ Check for American Language
This applies to institutions not located in the US. ChatGPT was developed and launched in America. Therefore, it has been trained in the American language. If your school is in the UK or other parts of the world, if you notice a paper is submitted in the American language, that should raise a red flag.
Overall, there may not be a full-proof method of figuring out whether a student has used ChatGPT. But if you combine the strategies mentioned above, you should be able to stop this unethical practice in your school. Some institutions have gone the extra mile by blocking the ChatGPT website on their networks.
In this part, we collect a set of frequently asked questions about Turnitin.
Unfortunately, Turnitin can’t detect the use of paraphrasing tools. The way Turnitin has been programmed, it’s only capable of noticing word or phrase matches. When a paraphrasing tool is used, it rewrites the content making it impossible for Turnitin to notice exact matches.
When some students discovered that Turnitin can catch plagiarism, they sought assistance from popular paraphrasing tools like Quillbot. If a paper has gone through Quillbot, Turnitin won’t detect plagiarism. Quillbot is one of the many tools that was developed to help learners beat Turnitin.
Here are some of the ways you can reduce similarity on Turnitin:
Even though Turnitin can check for matching text and AI-assisted writing, it can not detect everything. As a tool, it’s only limited to some functions.
The use of ChatGPT in the educational sector is surrounded by several ethical complications. However, thanks to Turnitin, instructors are now able to detect the use of ChatGPT in essays and other academic works. It may not be 100% accurate, but it’s adequately effective. Do you have any experience with using Turnitin to check for AI-assisted writing? Please let us know in the comments. And feel free to share this article with other people who are interested in Turnitin’s AI-detection capabilities.