Thank you for attending our AI Detection and Awareness workshops! We now offer a Turnitin AI detection tool for all campus instructors. Please note that while other options exist, our support will be focused on Turnitin.
Key Takeaways
Clearly Define AI Use for Your Students
- It’s essential to communicate your expectations regarding AI usage with your students.
- Recognize that the definition of “cheating” can vary among instructors and courses.
- Consider AI as a potential educational resource rather than a threat. For example, consider using AI as a possible tutor or study aid.
- Make sure to incorporate accessible AI policies within your syllabus.
Adapting Assignments for an AI-Integrated Environment
- Encourage students to reflect personally as they interact with AI tools.
- Request students to provide documentary artifacts such as outlines, rough drafts, and bibliographies as part of their assignments.
Utilizing Turnitin in Canvas for AI Writing Detection
- Turnitin, integrated into Canvas, offers a comprehensive AI Writing Detection feature.
- Emphasize that all submissions through Turnitin undergo thorough AI content scanning.
- To access the AI indicator, refer to the Similarity Report (Instructions on how to find the Similarity Report can be found here):
- The AI indicator is not visible to students.
- AI detection only works well with lengthy prose.
- It may not provide reliable results for bullet points, brief responses, short essays, or poetry.
- Currently, AI detection exclusively supports English text.
- Remember that AI detection scores are not absolute; any concerning findings should be discussed with students.
AI Detection Technology Progress
- Be aware that AI detection technology continually advances to catch up with generative models.
- Follow the LTC Blog for updates!
Supplemental Resources:
- Presentation Slide Deck
- Recording of the Second Session
- UWW AI Syllabus Language Resource
- UWS Chapter 14 – Student Academic Disciplinary Procedures
11/6/23 Update
UW-System OPID Webinar Series: Safeguarding Our Students, Instructors, and Universities: Privacy, Security, Copyright, and Generative AI
As generative AI expands in capabilities, questions surrounding personal privacy, copyright, and intellectual property have risen in the academic community. The University of Wisconsin System recently hosted a panel discussion bringing together three system experts to discuss these issues. Timestamps from the webinar that discuss the highlighted points are included in parentheses.
Key Takeaways:
- The best approach to generative AI use in the classroom is communication of expectations. A syllabus policy that clearly describes appropriate use, if permitted, will increase responsible integration and decrease confusion. (17:12, 37:52)
- Homework or other assignments, once completed and submitted to an instructor for credit or grading, is protected under FERPA. In the context of using open source generative AI detectors, instructors should be cautious of submitting student work where privacy violations may occur. Students should be given an opportunity to consent to their work being submitted to public detection software. (14:10, 17:55)
- UW System is currently working to understand how generative AI intersects with our current practices and tools. They are working on reviewing policies, creating a guidance document, and researching potential enterprise solutions. (39:05)
- Instructors can require students to sign up for accounts with ChatGPT. (7:51)
Additional LTC Workshops
Incorporating AI into your Assessments
- Wednesday, September 27; 3:15 pm – 4:30 pm McGraw 19A / Hybrid
- Friday, October 13; 10:15 am – 11:30 am McGraw 19A / Hybrid
Use Cases and Opportunities for AI
- Wednesday, October 18; 3:15 pm – 4:30 pm McGraw 19A / Hybrid
- Wednesday, October 25; 3:15 pm – 4:30 pm McGraw 19A / Hybrid
For full descriptions of each workshop, please visit our previous post.