Since ChatGPT launched in 2022, many professionals and students have used the website to help write content or get ideas. The AI tool has gained popularity in making people work less while still having a final product to show. Though there are debates as to the ethics behind using this and other AI assistant tools, one of college students’ main concerns is whether the content this website produces can be detected by plagiarism and AI checkers. Websites such as Turnitin and others have long been calling out plagiarism in students’ work, but can Turnitin detect ChatGPT? How exactly are students even using the website? And what do professors have to say about all of this?
Though many can agree that ChatGPT is an impressive creation, it has become a cause of concern among some professors and students alike. A survey conducted by Forbes showed that 89% of respondents use ChatGPT to help with homework assignments and at-home tests and quizzes, while 72% of college students believe the website should be banned. As for professors, 72% were concerned about how the tool can be used to aid in cheating, 34% believed it should be banned, but the remaining 66% were not in favor of banning the tool at all. The survey revealed that some educators are not necessarily against the website, but rather how students may abuse it and engage in academic misconduct. That’s where tools such as Turnitin come into play.
In response to the rise in ChaptGPT usage, Turnitin’s CEO Chris Caren wrote in April 2023 that his team, alongside feedback provided by educators, helped the company adjust quickly. Since OpenAI released GPT-3 in 2020, Turnitin has been researching and planning how to accurately show AI detection. In 2022, when ChatGPT was released to the public, Turnitin acted fast by accelerating their research and development, until they eventually launched their AI detection tool in April 2023. The website checks for the same things as before (originality and similarity), but now also includes a percentage for suspected AI content.
Caren states that the Turnitin team will keep adjusting as AI writing tools become more advanced. So in other words, yes, Turnitin can tell professors when a student has likely used ChatGPT. The company is also focused on ensuring that its detection tool does not give out false positives, as they are aware of the negative effects this might have. This is especially important since many universities have adopted new AI policies that might greatly affect students if it is believed that they fabricated their work using ChatGPT or other tools.
The University of Missouri and Harvard University are some of the schools that have issued statements on ChatGPT usage. The University of Missouri implied that using AI tools for work is a problem if it is used without the professor’s permission. However, if the professor grants permission, and the website is used in the conditions given, then the use of ChatGPT is allowed. Harvard stated that it “supports responsible experimentation with generative AI tools” but warns about the mistakes AI tends to make, the malware issues that might come with such websites, and the importance of consulting with professors before using the tool for academic works.
If you want to see Turnitin’s detection system in action, professor and AI enthusiast Elise Victor took to TikTok to share what the new AI detection feature looks like for teachers. In her TikTok, she shares how the detection tool comes with a warning that states, “Percentages may not indicate academic misconduct. Review required.” It also highlights parts of a student’s paper that the tool detects were written by AI.
If you insist on using ChatGPT and your school uses an AI detection tool, try to use the website responsibly. This can mean using AI to draft an essay outline rather than the essay itself, as one example. It might also be helpful for you to check your schools’ AI policies and email your professors beforehand to decide if it’s even worth using the website. Do whatever is necessary (and ethical) to prevent sanctioning from schools.