If you are a teacher or a professor wondering whether TurnItIn can detect ChatGPT, you’re not the only one. ChatGPT and other AI assistants have made it much easier to generate ideas, keywords and sentences for your work.
These AI assistants train on the data available on the internet, including scientific journals, articles, and original novels.
However, these tools have led to another issue — plagiarism in academic writing, especially in schools. This led many schools to ban ChatGPT. Although there are several tools to check paragraphs written by ChatGPT and other AI Assistants, TurnItIn is the one that is still leading the race.
In this article, we’ll deal with various questions regarding TurnItIn’s ChatGPT detection abilities.
What is TurnItIn?
Turnitin has been in the education sector since 1998. It started as a peer-review application to help students get feedback from their friends. The website was launched in 2000, and in 2007, the company began working with CrossRef to develop a massive content database to check scholarly articles for similarities.
After much time in the education sector acquiring firms and doing other things, TurnItIn began building AI-powered tools for further gatekeeping plagiarism and cheating in original essays and articles.
After GPT 3 was released in 2020, TurnItIn focused on AI writing detection.
In April 2023, after about three years of research, TurnItIn previewed the AI detection capability to institutions and educators.
Also read: How to connect ChatGPT to the internet?
Can TurnItIn detect ChatGPT?
Yes, TurnItIn can also detect ChatGPT’s texts and texts from other AI assistants.
The educator will upload the text or paragraph in the AI detector, and the detector will highlight the exact phrases that were copied from ChatGPT. One thing to note is that the AI detector will only work when the text is at least 300 words.
However, this information will not be shared with the student, and it’s now up to the educator to take further action.
How does TurnItIn detect ChatGPT?
AI chatbots owe their effectiveness to utilising expansive language models (LLMs). These machine learning models possess the capability to produce text that is grammatically correct and reads as if crafted by a human author.
Historically, TurnItIn was limited to conventional methods for detecting similarities in works. It maintained a comprehensive database of text drawn from various sources to accomplish this. Whenever an educator uploaded an assignment or document, TurnItIn cross-referenced this database to identify potential overlaps. if the submission aligned with an existing source, the software would highlight the matching portion and alert the educator.
However, that being said, ChatGPt and similar AI writing tools could easily circumvent TurnItIn’s basic plagiarism detection. Large language models don’t merely replicate existing content; they can generate entirely new text that has never existed. To address this vulnerability, TurnItIn introduced an AI writing tool in 2023.
Large language models like ChatGPT generate text in a highly systematic manner. Specifically, they employ a probabilistic approach to select words and construct sentences. In contrast, human writing lacks predictability. Consequently, even though ChatGPT was trained on text created by humans, its results can be readily distinguished from authentic human writing.
Apart from ChatGPT, the Turnitin AI flagging tool also effectively identifies AI-generated text within samples produced by Bing Chat.
Also read: Can you get banned from ChatGPT?
