Generative AI Guidance

Crafting your syllabus policy

No matter your stance on AI tools, the key is to clearly communicate your policies to your students.

Your policy should be based on what you decide is best for the learning objectives you’ve set out. You could choose from among the following, or you could combine some of these options For example you might disallow any use for the first half of the semester, then introduce use of it later. Or you might allow it for one or two assignments if students turn in both their work and work edited by generative AI as a comparison study. 

• Banned.  You may find any use at all to be too risky.  

  • A good fit when you want students to practice working independently, or practice lower-order writing skills like basic sentence composition and grammar. 

  • Difficult to enforce, may raise subsequent issues with trust and fairness.

• Limited.  You may want students to use AI in a regulated way. Some valid uses of AI might be: to help generate ideas, to improve grammar, or as a starting point for more in-depth research.  CELT’s document with sample statements  for syllabi also includes links to tips on how to cite generative AI. A good fit when:

• • you want students to focus on practicing higher order thinking and writing. 

  • you are ready to provide guidance to students about using generative AI critically.  

  • you want to encourage your students to reflect on the use of generative AI. 

• Allowed.  You may find it simplest to assume ubiquitous use of generative AI, and grade accordingly. A good fit when:

• • you are confident your students already understand all privacy concerns as well as how to use AI critically.

This is a collection of policies at other institutions. 

Crafting your syllabus and class plan

First, be transparent about your policy and provide an explicit statement about generative AI in your syllabus. CELT has developed sample statements to assist you in crafting one that works for your syllabus. You can find these sample statements above available to download or you can also find them at the bottom of this page in html format to copy and paste.

Second, have an honest discussion with your students about generative AI. Concerns specific to academia that you might discuss with students include the following, and include arguments both for and against student use of AI.

Optionally, adding a short bumper video to your course Blackboard that explains your rationale will serve two purposes: a) humanize your class and b) help students understand how generative AI could affect their learning outcomes. Michelle Pacansky-Brock has some great examples of brief humanizing videos. CELT also has several resources available if you need help creating a video.

Arguments against academic use of AI: 

• There are serious issues of student privacy. When students input their personal information or intellectual property into these proprietary tools, they lose control of that data. 

• Perhaps students will not develop strong research skills if they rely on a single tool, especially since generative AI sometimes makes up citations or facts, known also as “hallucinating”.  

• There are concerns that students will not complete their own work, and therefore will lose out on their own academic journey, for example, that they will have weak writing skills. 

Arguments for the academic use of AI: 

• Generative AI can support students with different learning needs. For example, ESL students may be able to use it to check grammar and formatting, as they would with a writing center tutor. 

• Concerns that if students do not practice using AI and reflect deeply on it they will not enter their post-academic careers fully prepared for using a ubiquitously available tool critically and responsibly. 

• Generative AI may be most similar to a calculator: if students use it to perfect lower order writing, they can then focus on more complex and interesting reflection and analysis. 

Can you win a technological surveillance battle if you ban generative AI?

The honest answer is: probably not, at least not in a majority of cases. At UAS, SafeAssign is our built-in automated tool to look for certain types of plagiarism. At this time SafeAssign will not flag anything written by any of the major generative AI. Other tools, such as ZeroGPT and Turnitin that have been specifically built to “catch” AI have also failed to do so reliably. (2) 

As anti-cheating tools evolve, generative AI likely will as well, so it’s best to assume that while they may provide an indication of some assignments to take a closer look at, we will need to rely also on further investigation of individual cases. Thus far, all major anti-cheating tools have provided significant numbers of both false negatives and false positives. Accusations of academic dishonesty in this regard should thus be considered with gravity. 

What can you do instead?

Updating your course design may be more effective than anti-cheating technology for preventing academic dishonesty. (3) 

Here are four suggestions for assignment modification that you can consider. Any of these will help minimize students’ over-reliance on generative AI, or provide structure for them to use generative AI reflectively and not just as a crutch. 

• Have students use alternatives to text. For example, oral exams or  VoiceThread  assignments. 

• Get samples of your students’ writing throughout the semester. This will give you a benchmark for comparison if you suspect unsanctioned use of generative AI. In addition, providing students with opportunities to turn in several low-stakes assignments is a tried and tested method for decreasing their inclination to cheat. 

• Choose thought-provoking prompts that encourage students to relate course concepts to their life experiences. If students understand how they will be able to use the product of assignment for their professional portfolio or later in their own academic journey, they are more likely to engage with the assignment (whether or not they get assistance from generative AI). 

• If you do suspect students are using AI, you can build in quiz questions that require students to recall what they have written about on past assignments. 

If you’d like to explore more options and concerns for incorporating generative AI into your classroom check out Leon Furze’s Teaching AI Ethics  and  Practical Strategies for ChatGPT in Education.

Finally, try putting the written assignments from your class into ChatGPT, Gemini, and other generative AI programs and seeing what output you get!

Sample statements for your syllabus

For courses that anticipate little or no AI use:

Generative AI Not Permitted:

Generative AI Permitted with Citation:

Generative AI Acceptable on Certain Assignments Only:

Example of instructions for permitting Generative AI, including wording when it is only permitted for a specific project

In-text citations

1. One-Third of College Students Used ChatGPT for Schoolwork During the 2022-23 Academic Year - Intelligent. (2023, June 9). Retrieved June 22, 2023, from Intelligent website:https://www.intelligent.com/one-third-of-college-students-used-chatgpt-for-schoolwork-during-the-2022-23-academic-year/

2. Can SafeAssign Detect ChatGPT? (Fast facts) - Technology Org. Retrieved June 22, 2023, from Technology Org website:https://www.technology.org/how-and-why/can-safeassign-detect-chatgpt/ 
   

3. Rettinger, David. “Show Students You Care About Their Learning—They May Cheat Less.” The Faculty Lounge (Harvard Business Publishing). May 3, 2022.

Recommended further reading

Condensed list of faculty advice (link is external)  from Inside Higher Ed, Jan. 12, 2023

Castañeda, L., & Selwyn, N. (2018). More than tools? Making sense of the ongoing digitizations of higher education. International Journal of Educational Technology in Higher Education, 15(22). 

Lasota, Dan. (2023, May 19). Dealing with AI generated academic dishonesty from a policy and teaching perspective – iTeachU. Retrieved May 24, 2023, from Uaf.edu website.

Promoting Academic Integrity in Your Course | Center for Teaching Innovation. (2021). Retrieved May 24, 2023, from Cornell.edu website.

ChatGPT and Good Intentions in Higher Ed. (2022, December 30). Retrieved June 1, 2023, from Is a Liminal Space website.

Artificial Intelligence in Education | Center for Innovative Teaching and Learning | Northern Illinois University. (2023). Retrieved June 12, 2023, from Northern Illinois University website.