Turnitin expands AI tools to combat grading biases in education

Turnitin has expanded its technological offerings aimed at enhancing academic integrity and improving learning outcomes for educators and students.

The new features, which are planned to assist teachers in the upcoming school year, utilise advanced AI capabilities to ensure fairer grading practices and improved detection of AI-generated content in student submissions.

The company introduced the Paper to Digital add-on, which leverages handwriting recognition and AI-assisted grouping to facilitate more consistent and equitable grading of paper-based assessments. The tool is designed to eliminate unintended biases, such as those arising from the alphabetical ordering of student submissions in learning management systems (LMS). In relation to this, a preprint paper from University of Michigan researchers has highlighted a significant grading bias favouring students whose surnames appear earlier in the alphabet.

Turnitin says its new tool aims to combat such biases by implementing horizontal, or question-by-question, grading. Chief Product Officer Annie Chechitelli noted, "This school year, AI will likely be in every classroom. Yet, there's still a disconnect between students and educators about what constitutes acceptable generative AI use and access to technology to support learning. Turnitin is helping to address these challenges by providing teachers and administrators with new offerings to adapt to AI's ever-evolving role in education, and supporting ethical AI usage for students."

In addition to facilitating fairer grading practices, the Paper to Digital add-on is also expected to streamline grading processes, enabling educators to provide faster feedback. This enhanced feedback mechanism integrates directly into existing LMS platforms, aiming to provide timely and customised student support.

Another significant update from Turnitin includes the enhancement of its Similarity Report. Initially launched in April 2023, Turnitin's AI writing detection feature was developed to help educators identify AI-generated content in student submissions. Based on feedback from educators and administrators, Turnitin has now refined this feature to offer a more intuitive and categorised user interface. The enhanced Similarity Report is designed to help educators pinpoint unoriginal or improperly cited materials, facilitating more effective discussions on citation practices and plagiarism prevention.

This update comes on the heels of Turnitin's recent introduction of an AI paraphrasing detection feature. Integrated within Turnitin's originality tools, this feature assists educators and publishers in identifying when AI paraphrasing tools have been used to adapt AI-generated text, thereby avoiding detection mechanisms. This tool is part of Turnitin's broader strategy to provide comprehensive support for maintaining academic integrity in the face of rapidly advancing AI technologies.

The expansion of Turnitin's technological suite comes at a time when AI adoption in education is growing, particularly among students. According to a study conducted by Tyton Partners, 59% of students regularly use generative AI tools, compared to approximately 40% of instructors and administrators. This discrepancy underscores the need for educational tools that support ethical AI usage and aid educators in adapting to these technological advancements.

Turnitin’s product portfolio and its recent enhancements have received widespread recognition. The company has been nominated for the 2024 Bett Awards and has won several accolades, including the 2024 EdTech Breakthrough Awards, 2023 Tech & Learning Awards, and 2024 GSV 150 Awards.