You are here

AI and Academic Integrity

Generative AI applications, such as ChatGPT, are widely available and are rapidly being adopted to maximize productivity. They can be used in a variety of applications, including text generation, content summarization, text to image, text to video, and multimedia design. 

In higher education, generative AI has raised concerns about academic integrity. However, it also offers opportunities to enhance teaching and learning. 

AI Detection

A common response among faculty is to rely on AI detectors to prohibit AI usage in their classes. For instance, Turnitin, a paper submission platform supported by our office, now offers AI writing detection. Turnitin’s AI writing detection and similar services use generative AI to detect patterns of writing commonly produced by generative AI. They report the results in the form of a score that predicts the probability of AI-generated material.

However, AI detection services have been found to be highly unreliable. Both Turnitin and OpenAI (the creator of ChatGPT) have acknowledged that AI detection is subject to serious inaccuracies and limitations. The following is a list of some of the most common problems with AI detection.

  • The detection algorithms often generate false positives. 
  • These tools can be circumvented by AI spoofers (e.g., undetectable.ai) that can make AI-generated material virtually undetectable.
  • The tools cannot provide material evidence of AI plagiarism, only a score, which can be inaccurate and disputed.
  • As AI is integrated into more and more applications commonly used by students (e.g., MS Office, email applications, and search engines), AI detection will become even more difficult to perform over time.

Therefore, reliance on AI detection services is not advisable. 

With the understanding that AI detection scores are a starting point rather than an endpoint, Turnitin has developed a number of helpful resources to assist faculty with engaging students in discussing suspected cases of generative AI use in their classes. These resources include:

To see all of Turnitin’s instructor resources for AI writing detection, visit their resource hub on academic integrity in the age of AI.

AI Prevention or Adoption

There are more effective ways to prevent the use of AI in the classroom. However, faculty should weigh the advantages of using AI before deciding to prohibit it. Teaching students to use these tools effectively and ethically in classroom activities and assessments may benefit them in the workforce. Additionally, there are many helpful AI tools that can assist faculty with teaching and learning. Therefore, faculty should carefully consider whether to choose the path of AI prevention or adoption. These topics are addressed on our Generative AI and Authentic Assessments page.

Syllabus Statement on AI

Whatever approach faculty decide to take, they should be transparent and clear about their policy on AI usage by providing a statement on AI in their syllabus. The policy should state what usage is or isn’t allowed and what are the consequences for violating the policy. If AI usage is allowed, faculty must consider privacy concerns as well. 

Below are sample statements prohibiting or encouraging the use of AI.

Sample Statement Prohibiting Use of AI-generated Material

All work submitted for this course must be your own original production. Use of artificial intelligence (AI) to generate material is strictly prohibited. Submission of material generated by AI will be considered a violation of the University’s academic integrity policy. This includes AI-generated material that you have summarized or edited.

Sample Statement Encouraging Use of AI-generated Material

You are responsible for the content of any work submitted for this course. Use of artificial intelligence (AI) to generate a first draft of an assignment is permitted, but you must review and revise any AI-generated material before submission. In addition, you are required to be transparent about your use of AI-generated material by following citation and documentation guidelines provided in this course. Keep in mind that while AI generators can be useful tools, they are prone to factual errors, incorrect or fabricated citations, and misinterpretations of abstract concepts. Utilize them with caution.

The following links provide additional examples, templates, and resources for crafting a syllabus statement on AI.

Citing & Documenting AI-generated Material

Generative AI raises concerns about intellectual property and attribution. While the answers to these questions are still being worked out, it’s advisable to require students to be transparent about when and how they use AI-generated material in their work. Some tentative guidelines for how to cite and document AI-generated material can be found on these pages: