Artificial Intelligence and Academic Integrity


ACADEMIC INTEGRITY

The following pages are part of the Academic Integrity toolkit entries.

You can jump around or review each page in sequence by following the hyperlinks:

  1. Academic Integrity Section Overview
  2. Promoting Academic Honesty
  3. Investigating Suspected Academic Dishonesty
  4. Academic Integrity in Online Environments
  5. Artificial Intelligence and Academic Integrity (This page)
  6. Sample Artificial Intelligence Syllabus Statements
  7. Additional Resources

What do I need to know about academic integrity and artificial intelligence?

Like all forms of academic dishonesty, it is critical that faculty assess their ethical stance on academic misconduct with artificial intelligence (AI). In order to gain a better understanding of AI, Instructional Technology has created a webpage and several workshops that you can sign up for to learn more. You can also review their presentation, ChatGPT and AI in Education Links to an external site. or download the slides Download download the slides. Review this memo and a FAQ guiding document to learn more about guidance CSUN is providing to faculty on this topic. 

Suggestions 

Be proactive about addressing the use of AI in your course.

  • Be explicit in your expectations on the use of AI as you would be with your classroom policy on academic integrity. Include this in your syllabus. 
  • Educate your students on what is acceptable and not, as it pertains to both AI and academic integrity. You may need to have this conversation multiple times. Work with the library, instructional technologists, faculty development, and student conduct to get guidance and tips.
  • Programs that have been used in the past to help student’s edit their papers, such as Grammarly, are now being flagged as AI generated content. It is important that you determine if the use of proofreading applications is permissible in your classroom. It is worth asking yourself if utilizing proofreading applications was an acceptable practice in the past, then what makes this any different?
  • Possible mitigation strategies:
    • Show students how to use AI ethically and responsibly
    • Students must cite when they use AI
    • Allow students to rewrite in their own voice when they are caught. Give them a chance to correct it.
    • Modify assignment prompts to focus on critical thinking, personal opinions, and life experiences. 
    • Use a transparent assignment design approach when planning an assignment that uses AI. FacDev offers the Get Up to Speed with Transparency in Teaching online course and the Transparent Assignments series, which provide you with guidance and support in designing transparent assignments. The Ohio State University has a teaching guide Links to an external site., which can help you design a transparent assignment that uses AI. 

How reliable are AI detection scores?

Instructional Technology’s presentation, ChatGPT and Plagiarism Detection Tools Links to an external site., delves deeper into the AI issues relevant to higher education. There are issues with AI detection scores. They are not reliable or 100% accurate. 

Turnitin is a plagiarism detection software that is automatically embedded as a Canvas activity option for submitting course papers. Turnitin includes a disclaimer about anything under 20%:

“We learned that our AI writing detection scores under 20% have a higher incidence of false positives. This is inconsistent behavior, and we will continue to test to understand the root cause. In order to reduce the likelihood of misinterpretation, we have updated the AI indicator button in the Similarity Report to contain an asterisk for percentages less than 20% to call attention to the fact that the score is less reliable.”

You can review more on their webpage, How do Turnitin’s AI writing detection capabilities work? Links to an external site.

In addition to Turnitin, there are other plagiarism detection tools available including GPTZero Links to an external site., however, all AI detection tools on the market have the same problems with providing false positives and are far from being 100% accurate.

What is being flagged in AI detection tools?

Original writing has been flagged as AI-generated, which highlights the fact that detection scores are not 100% accurate. In addition to detecting the use of Chat-GPT, AI detection tools are also flagging the use of applications that have been used for years. Some of these programs include Grammarly and Google Translate.

Responding to incidents that go against the AI statement policy in your syllabus 

Like all incidents of academic dishonesty, the most important thing you can do is speak to the student when you discover it or are suspicious. Review the guidelines on approaching a student if you suspect academic dishonesty.

Read the content. The content may miss the mark. In certain instances, it is blatantly wrong and not in line with the assignment. Gathering evidence as it pertains to AI is like putting together pieces of a puzzle. Some pieces are not as prominent or weigh as strongly as others. For example, comparing tone in assignments is not as strong as erroneous content. 

Keep in mind the burden of proof for responsibility falls on the university. Relying solely on an AI detection score is not sufficient evidence when a student is facing disciplinary action from the OSCED.