New AI detection feature via Turnitin

Dr. Rachelle Dené Poth
6 min readApr 7, 2023

--

Over the last couple of months, it’s unbelievable how much has changed in the world of artificial intelligence (AI). For the last couple of years, I have spent time researching, writing about, presenting on, and teaching about AI in my classroom and continuing to learn as much as I possibly can each year. When ChatGPT became available to the public in December, I didn’t hesitate to dive in and explore exactly what it was capable of doing. I even co-wrote an article about its potential impact on not just education but other areas of work.

After about 3 hours, I was pretty impressed with what I was seeing. I was also a little bit concerned about what it could be used for, whether by educators, students, or anybody for that matter. ChatGPT and the evolving AI technologies are not something that will impact educators and students, it is something that will impact many areas of the workforce and jobs that are out there. A big concern that has come up during this ChatGPT exploration is plagiarism. How will educators be able to tell when students may be using ChatGPT to complete their assignments, especially with something evolving as fast as this technology? Will students begin to rely on this technology and as a result, lose the opportunity to develop essential skills on their own? What policies may need to be in place? Is it plagiarism if it is not citing a specific source? These are some of the many questions on the minds of educators and parents.

Plagiarism detection

There have been plagiarism checkers available for years and many educators and educational institutions use them. In my own experience as a doctoral student now, and when I worked on my Master’s degree five years ago, assignment submissions go through a plagiarism checker instantly. Turnitin is one that has been used. Honestly, there have been times when I’ve been shocked to see the percentage that comes back stating the likelihood of plagiarism, knowing that I did not plagiarize. I work hard to cite my sources but apparently, something that I did was not completely accurate and I used it as a way to evaluate my writing and improve my citations. Now enter ChatGPT.

Three years ago, I helped to do some research for a blog post about GPT-3 and I was impressed with what it was capable of back in 2020. Fast forward to early December 2022, working with and exploring ChatGPT, I was amazed at how much it had evolved since 2020 and the number of ways that you could use it. It didn’t take long for educators and lots of people to express concern about what the impact of this type of powerful technology would have on student learning.

Would it lead students to lose learning opportunities?

Would they rely on everything that they found by using this, and lack skills that they need to be prepared for the future?

What should educators do when it comes to teaching about these technologies in the classroom?

How can you tell whether or not a student has used ChatGPT to complete an assignment or project and how do you handle it?

Learning opportunities

AI writing and the use of ChatGPT and other AI writing tools are increasingly common not only in business. Students are using it and know about it. There are high schools banning its use. Educators are worried it will erode writing skills. But is banning it the solution? Detecting the presence of artificial intelligence in a student’s writing is helpful, but I hope that it is focused not on catching them doing something wrong but rather to use it to improve writing. We should use it to start conversations with students.

Turnitin has been developing writing and feedback tools for educators for 25 years and has continued to refine the ways that teachers give feedback to students. Turnitin commissioned Atomik Research to conduct an online survey of 1,011 parents and/or guardians of high school students in the United States to understand their perceptions of AI writing tools like ChatGPT and Google Bard. Interestingly, 81% of respondents believe that teachers should use technology tools to detect when something has been written by AI to check homework or test answers to cut down on cheating.

On Tuesday, Turnitin released a new AI detection feature that has the ability to detect AI with 98% accuracy. They are also continuing to monitor the detection settings to watch for false positives. The new AI detection feature works like similarity checking. If something is noted, then teachers can look at the writing and provide feedback to students. It becomes a conversation and a learning opportunity for teachers and students about how to evaluate information and also for properly citing sources that have been used. The software should not be viewed as something that is set to accuse students of plagiarizing.

[image via Turnitin]

Also helpful is the sidebar with resources for educators to explore.

[Similarity report via the AI writing feature]

As educators, we have a responsibility to help our students to develop skills to navigate all of these changes that are happening in the world of education, and that may impact them in their future line of work.

Lessons to learn

I see this new feature as a way to provide information for teachers so they have data for analyzing student writing and providing further instruction and feedback, not as a tool for accusing students of misconduct. We want students to understand the importance of academic integrity while also learning about the tools available, including ChatGPT and other AI technologies. It is a way to also build literacy skills and stress the importance of evaluating our sources and checking information for accuracy.

Now what?

I recommend following the conversation about ChatGPT in the variety of educator spaces and communities. As for tools that detect plagiarism, explore the resources available to assist educators in integrating this new technology into their classrooms. The Turnitin page features a glossary of AI terms, which provides a useful reference guide for educators who are not yet familiar with the technical vocabulary associated with this technology. Additionally, there is a guide on updating an academic integrity policy in the age of AI, which is essential for ensuring that educators have a clear understanding of the ethical considerations surrounding AI-generated text.

And the best advice I can offer is to dive into trying out ChatGPT. Think about questions that you ask your students, projects that you have assigned, assessments that you’ve created, and lesson plans that you’ve written, and put all of those in as prompts. See what it generates and then regenerate the response a few times and look for commonalities. I test this out to see how similar the answers are and also check it for accuracy. I have explored this with my students as well because we know that not everything that we find on the Internet is accurate, which is why we have to embrace this and help students to learn to be digitally literate and model this practice for them.

This post was in collaboration with Pando PR. Opinions expressed are my own.

**Interested in writing a guest blog for my site? Would love to share your ideas! Submit your post here. Looking for a new book to read? Find these available at bit.ly/Pothbooks

************ Also check out my THRIVEinEDU Podcast Here!

Join my weekly show on Fridays at 6 pm ET THRIVEinEDU on Facebook. Join the group here

Originally published at http://rdene915.com on April 7, 2023.

--

--

Dr. Rachelle Dené Poth

I am a Spanish and STEAM Emerging Tech Teacher, Attorney, Author and Blogger, Learning Enthusiast and EdTech Consultant