Faculty Guide: Assessing Turnitin AI Writing Detection Reports for Academic Integrity

Written by Chris Bryson, Ph.D., Executive Director for Academic Initiatives and Integrity

Turnitin is now available to all UARK instructors. One component of this software will improve an instructor’s ability to assess whether the written work submitted by a student contains text generated by artificial intelligence (AI). The use of AI tools can be considered academic dishonesty if a student has not received permission to use such tools from an instructor in advance. For this reason, it is critical that UARK instructors understand how to utilize Turnitin’s AI Writing Detection report in determining whether they should submit a report to the Office of Academic Initiatives and Integrity. Like writing flagged by plagiarism software, a percentage of flagged text within an AI Writing Detection report should not automatically result in a report of academic misconduct. Rather, this alert should call for further review by the instructor, so here is what you need to know.

Turnitin’s Writing Detection software is a strong indicator that AI-text may be present. Weber-Wulff et al. (2023) found that Turnitin outperformed 13 other AI-text detection tools and identified 81% of AI-generated text in various forms. The report UARK Instructors have access to highlights text likely to be generated by AI in blue, and text likely to be re-paraphrased by AI in purple. If such flag(s) are present within a report, concerned faculty should ensure that the work submitted includes a minimum of 300 words and the percentage flagged is over 20%. Under these conditions, Turnitin has claimed a false positive rate of less than 1%, which has been validated across multiple neutral third-party studies (Walters, 2023; Weber-Wulff et al., 2023). At this point, instructors should further evaluate the work to see if it includes common factors associated with the use of AI tools. Such factors would include, but are not limited to, the following:

  • The work submitted does not address the assignment’s question
  • The work submitted uses concepts that were not addressed within the course content
  • Previous work submitted by the student is much different than the work in question
  • Clear statements within the text that reference the use of a large language model

Instructors should consider reporting the issue for further evaluation by the Academic Integrity Monitor (AIM) in their college if one or more of these issues are present. It should be noted that instructors are welcome to discuss their concerns with students to learn more about their writing process prior to reporting if they wish; however, the academic integrity policy approved by UARK’s faculty senate requires all instructors to report issues of academic misconduct they suspect occurred. If any of the issues above exist, along with flagged text within an AI Writing Detection report, it is well within the scope of an instructor’s duty to report this issue to be further investigated by an AIM. Please know that OAII staff are available to consult with an instructor about a particular situation by phone at 575-5229 or at honesty@uark.edu, or they can reach out directly to their AIM.

To learn more about Turnitin’s AI Writing Detection software, please see Turnitin’s white paper and FAQs.