Generative AI: Academic Integrity

Considerations for Assessments and Academic Integrity

Essential Considerations for Addressing the Possibility of AI-Driven Cheating, Part 1 (Article)
Torrey Trust PhD, Faculty Focus
Discusses six considerations related to AI and cheating:

  1. Banning AI chatbots can exacerbate the digital divide.
  2. Banning the use of technology for exams can create an inaccessible, discriminatory learning experience.
  3. AI text detectors are not meant to be used to catch students cheating.
  4. Redesigning academic integrity statements is essential.
  5. Students need opportunities to learn with and about AI.
  6. Redesigning assignments can reduce the potential for cheating with AI.

Considerations for Redesigning for Assignments

Trust (2023) describes how students are most likely to cheat in these situations:

  1. Focus is on grades more than learning.
  2. Increased stress, pressure and anxiety for student (e.g. use of high-stakes assessments like one-shot exams).
  3. Lack of focus on academic integrity, relationship building, and trust (e.g. little effort to establish norms and set expectations about the importance and value of academic honesty).
  4. Learning material is not considered relevant or valuable by students.

In his article, he lays out a TRUST model:

Transparency – be clear about what the assignment is about, why they are doing it, and the steps they need to take to complete it.

Real World Applications – have assignments relate or contribute to the real world.

Universal Design for Learning (UDL) – keep the three principles of UDL in mind when designing assignments: try to provide options for Multiple Means of Engagement, Multiple Means of Action and Expression, and Multiple Means of Representation.

Social Knowledge Construction – encourage social learning in which learners co-construct and share understanding.

Trial and Error – give students opportunity to learn through productive failure by including opportunities to re-submit and offering low-stakes quizzes that can be taken multiple times. This takes away some of the pressure of one-shot, high-stakes assessments.

Setting Expectations for AI Use in Your Course

Even if you don’t use generative AI in your learning activities, you will need to set expectations about student use of AI in your course.

1. Confirm your course policy on Generative AI Use

Be sure to consult with your colleagues and program coordinator to ensure that you are aligned with the general approach of your program.

This curated list of Classroom Policies for AI Generative Tools from various higher education institutions can provide you with ideas.

Algonquin College defines the following scenarios:

  1. Use Prohibited
  2. Some Use Permitted
  3. Sourcing permitted, with citations
  4. For research purposes, with citations
  5. Full use, with citations
  6. Unrestricted Use

2. Include statements within your Weekly Schedule (or Course Section Information – CSIs) and Assignments

Algonquin College has drafted some example statements that you can use in your Weekly Schedules.
You may also consider having students acknowledge their use of Generative AI in their assignments. View Student Statements: Acknowledgement of AI Use

3. Discuss what AI-related academic dishonesty means in your course

Be sure to discuss with your students what appropriate and inappropriate use of generative AI looks like in your course. You can confirm, in reference to policy AA48: Academic Integrity, that using Generative AI to write or create any aspect of an assignment without proper citation is considered academic dishonesty. Also, students are not allowed to complete quizzes, tests, or exams using ChatGPT or other generative AI tools.

4. Show students how to properly cite their use of Generative AI Tools

It is a good idea not to assume that your students know how to cite their use of Generative AI. Guide your students in how to include AI-generated content into their assignments and to properly cite it.

Citing AI Content by Citation Style (Algonquin College Library)

Detection of AI Use

The Office of Academic Integrity shares tips for how to identify and fairly investigate AI use.

Faculty Guide: Fairly Investigating Misuse of GenAI (pdf)

AI Detector Tools

Turnitin (TII) had made available its AI Detector Tool freely available until the end of December, 2023. It was a “probability of the presence of writing created by Generative AI” score that was part of Turnitin’s Feedback Studio. Turnitin is an assignment feature integrated into Algonquin’s Brightspace platform.

During the F23 semester, Algonquin reviewed the tool to confirm whether it should be purchased. The decision was made not to purchase it for these key reasons:

As large language models become more sophisticated, it appears the feature will become less capable of identifying AI-generated text, combined with an increasing false positive rate.

Turnitin is not the only AI detector that is unable to accurately and reliably identify AI generated text. To our knowledge, there are no products on the market that are regarded as being suitable for use in higher education. Notably, OpenAI, the creators of ChatGPT, have abandoned development on their own AI writing detection tool.

Furthermore, controlled studies have demonstrated that non-native English speakers are disproportionately impacted by false positive results.

Details of this decision was communicated via a memo from the Senior Vice-President, Academic on January 2, 2024.

Read Memo: Update on the Use of the Turnitin Detector – January 2, 2024

 

References

Coley, M. (2023, August 16). Guidance on AI detection and why we’re disabling Turnitin’s AI detector. Brightspace | Vanderbilt University. https://www.vanderbilt.edu/brightspace/2023/08/16/guidance-on-ai-detection-and-why-were-disabling-turnitins-ai-detector/#:~:text=After%20several%20months%20of%20using,tool%20for%20the%20foreseeable%20future.

Trust, T. (2023a, August 2). Essential considerations for addressing the possibility of AI-driven cheating, part 1. Faculty Focus | Higher Ed Teaching & Learning. https://www.facultyfocus.com/articles/teaching-with-technology-articles/essential-considerations-for-addressing-the-possibility-of-ai-driven-cheating-part-1/

Trust, T. (2023b, August 4). Essential considerations for addressing the possibility of AI-driven cheating, part 2. Faculty Focus | Higher Ed Teaching & Learning. https://www.facultyfocus.com/articles/teaching-with-technology-articles/essential-considerations-for-addressing-the-possibility-of-ai-driven-cheating-part-2/

University of Wisconsin – Green Bay (2023, May 2). Strategies for Creating “Generative AI-Resistant Assessments. https://blog.uwgb.edu/catl/strategies-for-creating-generative-ai-resistant-assessments/