Include information in the syllabus and on assignments, particularly if there is variability in when or how much AI use is allowable. As AI tools become increasingly embedded in existing technologies, students may be in situations where it is unclear how the use of the tool relates to your expectations of them. If the students are unsure, instruct them to contact you for clarification. 

Sample syllabus or assignment language: 

• *When content-generating AI is not allowed.This course assumes that all work submitted by a student will be generated independently by the student or as part of an assigned group. Any substantive portion of an assignment done by someone else, including AI-generated content, is not allowed and will be treated as academic misconduct.
• When content-generating AI is allowed with appropriate attribution. In all academic work, the ideas and contributions of others must be appropriately acknowledged. AI tools, including ChatGPT, are permitted in this course for certain assignments, and specific instructions will be included with each assignment. Identify any writing, text, or media AI generates when submitting work. Students should indicate how AI tools informed their process and the final product, including how AI-generated citations were validated. Failure to properly acknowledge the AI-generated contributions will be treated as academic misconduct.
• When content-generating AI use is allowed in limited instances. Students may use AI tools to help preparefor assignments and projects (e.g., to help with brainstorming, concept development, iterations of an idea, etc.). When submitting a final product for grading, students must indicate how AI tools informed their process and the final product, including how AI-generated citations were validated. Students are responsible for the accuracy of the generated content.
• When content-generating AI use is encouraged broadly.Students are encouraged to use AI tools in this course. When submitting work, students must clearly identify any writing, text, or media generated by AI. Students are responsible for the accuracy of the generated content. Failure to properly acknowledge the AI-generated contributions will be treated as academic misconduct.

*When content-generating AI is not allowed, consult Student Conduct Artificial Intelligence guidance.

To see the range of syllabus statements available by educators in navigating AI-generative tools, consult Lance Eaton, Director of Digital Pedagogy for College Unbound archive of sample AI policies.

AI-generative tools have brought to the forefront an examination of why we assign assignments to students. Consider these implications when designing assignments and assessments. 

• Clarify Objectives: Identify the knowledge, skills, and abilities that each assignment is helping learners develop. Consult CELT’s Writing Course Goals/Learning Objectives to connect your objectives with course content, assessments, and your teaching approach. 
• Scaffold the assignments. Build in scaffolding, requiring a proposal, outline, rough draft, and final draft as a part of the process. Plan Your Course worksheet (docx) can aid instructors in this process.
• Examine your course schedule and be flexible with due dates. Too many assignments, with too little turnaround time, can lead to students looking to take shortcuts, especially when coupled with stringent late policies. Use the Rice University Course Workload Estimator to assess the anticipated time investment of students and adjust due dates if needed.
• Partner with an Iowa State librarian. Invite an ISU librarian to discuss information literacy and proper citation of AI tools in your course. The specific citation will depend on the citation format (MLA, APA, Chicago, etc.) used in your course. For guidance, please get in touch with your subject librarian or email Jeff Kushkowski, Head of Instruction Services, at kushkows@iastate.edu.

Adapt existing assignments to deepen student reflection and have an appropriate depth of complexity and specificity. Consider how assignments that require concrete and factual information (easily obtained by AI-generative tools) might be redesigned to require more abstract and metacognitive thought and reflection on the part of the students. 

• Reflection: Require students to reflect on their research and writing process, what they learned from it, and how they would approach a similar task in the future.
• Mind Maps: Have students create a visual ‘mind map’ to illustrate the connections between ideas, concepts, approaches, or theories.
• Debates: Have students debate a major question or challenge. Even short debates can deepen learning and get students to look at topics from varied perspectives.
• Written work tied to specific course materials: Design writing prompts that require careful analysis of multiple course materials specific to your class.
• Required References: Alter assignments to require students to reference a certain number of current events or recent articles that would not yet be included in the AI’s database.

AI-generative tools can help learners generate ideas, identify gaps in thinking, and propose visual elements. These examples provide methods to incorporate AI-generative tools into your course assignments.

• AI Response Review: Identify a major question or challenge generative AI tools could write about. Ask students to create a rubric to assess the response, then review other students’ responses and rank them based on criteria.

• Reflect and Improve: Identify a major question or challenge in your discipline and prompt AI-generative tools to write about it. Ask students to reflect on the output (e.g., what is correct, what is incorrect, are there sections of the output where they don’t know if it is correct or incorrect, what should they look up elsewhere to verify, describe why they chose that resource as a reliable source of information, etc.). Have students edit the output using a ‘track changes’ feature to improve the response.

Note: When developing assignments and projects, it is essential to recognize that not all AI tools are designed for universal accessibility, which may limit their use for students.

If you are interested in experimenting with AI in your course, explore Understanding AI Writing Tools and their Uses for Teaching and Learning at UC Berkeley by Berkeley’s Center for Teaching & Learning.

Like using Turnitin software to detect plagiarism, there are tools designed to detect AI-generated content. However, there are significant limitations to using these tools, which are both technically and ethically complex. AI detectors are themselves AI tools and may indicate dishonesty when none exists. Further, many tools are more effective when analyzing large pieces of text rather than sentences or paragraphs, producing unacceptable false positives. Due to the unreliability, Open AI, which created Chat GPT, shut down its detector at the beginning of August due to low accuracy rates. 

Accusations of student use of AI tools, when not allowed, may create an adversarial relationship with students and have other unintended consequences. A blog post by University of Mississippi instructor Marc Watkins shows examples of output from multiple detection tools and their inaccuracy.

There are ways of identifying misconduct without utilizing AI detectors. These strategies include:

• Enter the assignment prompt into generative AI-tools to identify sample responses and gauge the level of complexity and similarity of the created response to the concerning student work. 
• Compare the submitted work to prior student assignments noting significant changes in tone and style of writing.
• Be mindful of AI “hallucinations” or “confabulations” in which generative-AI tools create realistic sounding information that is not accurate.
• Identify lack of citations, where these were required, or lack of accurate content matching citations provided. Some generative AI tools produce content that references “others,” “critics,” and “research,” without attribution.
• Identify references or topics that veer far outside of the class content or expectations for the assignment.

View Sarah Ruth Jacobs Inside Higher Education article for additional ways to make courses resistant to plagiarism.

Ultimately, as stated by Marc Watkins  “creating a culture of trust” is more important than “a rote and imperfect method of detecting cheating.” Review CELT’s achieving academic integrity page to reduce the risk of academic misconduct. Contact celt@iastate.edu for additional teaching resources or Assistant Dean and Director of Office of Student Conduct Sara Kellogg at skellogg@iastate.edu.