Have students use AI as a tutoring tool to help them prepare for tests by checking their understanding of course material.

Example: Open AI’s ChatGPT4.0 is now available for free and offers enhanced capabilities in understanding and discussing images, voice, and text in over 50 languages. Students could use tools like this to practice their speaking and writing skills (in English or in other languages) in real time. AI can provide customized, constructive feedback on aspects such as pronunciation, lexical choice, register, and tone. You might consider asking students to submit a transcript of their AI tutoring session as a formative assessment.

Ask students to evaluate AI-generated content and compare it to human-generated content to boost their AI literacy and critical thinking skills.

Example: Task students with identifying key ideas and main arguments in an AI-generated text and then ask them to find and cross-reference external source material to engage in lateral reading. Students could work in pairs to create an annotated list of sources that includes their thoughts on the motivation of the authors, the evidence and logic they use to build their argument, any biases they detect, as well as omissions and hallucinations they find in the AI-generated text.

If using AI does not align with your learning goals, try reframing your assignments to make it more difficult to do so.

Example: Consider incorporating visual elements (images, videos, etc) into assignments or discussions. This may deter students as they are more difficult to copy into AI tools. Ask students to analyze/interpret/provide context for the visual element.

Example: Incorporate something specific from a class discussion (i.e., “Analyze and critique a key argument from one of our class discussions, using evidence from the assigned readings and your own perspective to support your analysis.”) Please note that this strategy is not foolproof, as you typically cannot “outprompt” AI.

Example: You might try integrating your readings and requiring that they integrate, through direct quotes, those readings in student’ written responses. It takes a little bit more work to get AI to do this and might be a deterrent.

Example: Timed assignments may be helpful for short written responses or quizzes. The idea is that limited time may prevent students from accessing AI.

Design rubrics to evaluate AI generated content if you are concerned about students’ misuse.

Example: Some folks suggest emphasizing specificity and concision on your grading rubric (as both are problematic for AI). Some faculty basically tag portions of the writing with YAST (You Already Said This) and give point deductions, as AI responses are known to be repetitive.

Example: AI is good at summarizing, but not great at making connections. Typically AI does not do a good job of weaving together specific sources to form a coherent argument or conclusion. If you use a rubric, you could have a section on synthesis, for example, "articles are connected to one another and strengths/weaknesses and similarities/differences are summarized. A paper that gets full marks will show how all articles fit together into a coherent point."

Example: Some descriptors in the rubric could include:

• Unusual or overly-complex sentence construction: AI might sometimes generate sentences that are grammatically correct but unusual or overly complex, especially when it’s trying to generate text based on a mixture of styles or genres.
• Overuse of certain phrases or vocabulary: AI models can sometimes overuse certain phrases or vocabulary that it has learned during training. This is the result of the AI trying to fill up space with relevant keywords.
• Lack of depth/analysis: AI generated writing is better at static writing tasks than creative or analytical writing.
• Format and structure: AI-generated sentences are typically shorter and lack complexity, creativity, and varied sentence structures.

Review your course learning objectives and consider whether AI tools might support these at various stages of the learning process.

For instance, could AI be helpful as a beginning-of-process tool (i.e., at the brainstorming or drafting stage of a project?) or would it be more suitable as an end- of-process tool (i.e., at the revision or final editing stage?). Leon Furze takes a deep dive into this topic in The Myth of the AI First Draft. Your own approach will depend on the learning outcomes you want your students to achieve.

Many faculty are alarmed at potential violations of Academic Misconduct related to AI and feel concerned about what the future may hold as AI becomes more capable of mimicking human voice and decisions. What Steps Do I Take if I Suspect Academic Misconduct with AI?

The AI Guidelines and Resources Committee suggests extreme caution when considering AI detection tools. The bottom line is that no conclusive proof can be provided by an AI detection tool at this time. Given the easy access students have to generative AI tools, we find ourselves at a crossroads in education. Why TTU Recommends Caution with AI Detectors

Communicate with students about AI in the syllabus, in class during the first week of the semester, and again each time a task is assigned to remind them of your expectations for how they will use or not use AI in your course. The specific nature of your policy is up to you! As you think about what AI use you will (dis)allow, it can be extremely helpful to begin with the intended learning outcomes of your course and then consider how the use of AI might enhance, limit, or shift these.

Here’s a sample policy from a course that includes several smaller writing assignments. These instructions are included in each writing assignment to remind students of the syllabus policy discussed on the first day of class:

• "As you work on this assignment, you are welcome to use ChatGPT or another AI tool to assist you with brainstorming ideas, creating an outline, and drafting your first version of the text. In line with our course policy, up to 40% of your first draft may be generated by AI, and you must indicate these portions of the text by marking them in red."
• "During the peer review stage, AI use is prohibited, because it is important for your peers to hear your authentic voice and learn from your perspective: How did what they wrote “land” with you? What questions do you have after reading it? What would you suggest to improve upon the first draft? I have provided a rubric to guide your review, and you will be asked to insert comments into the text you are reviewing."
• "When it comes to the final draft of your paper, AI use is also prohibited. At this stage, you will have everything you need to produce a polished draft, including feedback from me and two of your peers. This policy supports one of the course's main learning objectives: To foster a community of human learners who support each other in developing critical thinking skills. That is why we agree to use AI in ways that enhance our efficiency without diminishing our ability to think independently and build connections with each other."

Everywhere that we turn, it seems like we hear about artificial intelligence and predictions about its impact. Dean Inara Scott at Oregon State University suggests although a “one size fits all” approach to AI is ill-advised, one universal directive is clear: Faculty cannot ignore AI. Our students are using it now and will be using it in their future careers. Perhaps the starting point is to assess your digital mindset (Neeley and Leonardi, 2022), in other words, to reflect on your attitudes and behaviors about AI and consider how you may need to stretch to see new ways of thinking about adopting AI and potential changes to your teaching. Here’s a simple plan adapted with permission from Dean Scott’s suggestions to our context here at Texas Tech:

1. Learn the basics. If you haven't already, spend time using some form of generative AI. Your discipline's tools may vary, but at a minimum, everyone can start with free versions of ChatGPT or Bing to get familiar with the technology. Does this sound overwhelming? Contact the TLPDC and ask for Lisa Low, our faculty fellow. Lisa would be glad to meet with you to have a hands-on look and conversation about your classes. There are also many articles online related to ChatGPT and generative AI in the classroom. Consider this flow chart developed by the AI Resources and Guidelines Committee or this overview and list of resources from the Oregon State University Center for Teaching and Learning.
2. Choose your AI policies. At Texas Tech, instructors set the AI policies for their classes based on their teaching philosophies and learning goals. You can find three recommended syllabus statements from the AI Resources and Guidelines Committee. Consider the ways in which your students might use AI. The Committee does not recommend prohibiting AI in your course as you are likely to spend too much time policing this policy and AI is now woven into almost all forms of web searching, word processing, spreadsheet generation, and presentations software. Regardless of what policy you select, talk to your students about your “whys.”
3. Consider Modifying or even Eliminating "Busy Work" Assignments. Students will be inclined to use AI to do work that they do not see the value of or don't have to engage with, which might include summarizing a reading, responding to simple discussion prompts, or the traditional 2–3-page essay assignments.
4. Review more difficult assignments: Are you testing critical human skills? Students will be inclined to use AI for high stakes assessments particularly if they aren’t confident and of course, we know that this is not a new practice, but just a new form of temptation). It is the instructor’s prerogative and disciplinary expertise that allow them to know what students need to learn but in our new AI environment, it’s worth considering what information is foundational and needs to memorized and what skills are essential.

Simple Ways to Get Started with AI by Suzanne Tapp

1. Communicate clearly your expectations regarding AI use - i.e. When is it (not) ok to use AI in this course? - in the syllabus and in conversation with students during the first week of class.
2. Help students connect course learning goals with their own sense of purpose. Why should students care about the knowledge and skills they will gain in this course, and why is it important that they participate actively in their own learning? How could AI tools support or hinder them in this process?
3. Integrate multiple modes of engagement to enhance writing assignments. Some examples are asynchronous annotations of written work, peer feedback rounds, short oral presentations, the collective development of mind maps, and other opportunities for students to demonstrate their engagement with course material.
4. Adjust assessments to reward the process of learning instead of just the final product. For example, a final essay assignment could be broken down into smaller tasks that reward effort throughout the semester and ensure a better final product:
   1. Week 7
      Asynchronous brainstorming on a collective document to gather topic ideas, research questions, and primary and secondary works to consider. (Students earn points for contributing to the document.)
   2. Week 8
      Research pairs are established, topics are selected, and research pairs submit a project plan that includes a timeline for completion. (Students earn points for submitting a plan.)
   3. Week 11
      Research pairs submit a research poster draft and receive feedback from one other research pair via a rubric. (Students earn points for submitting a draft AND for providing feedback.)
   4. Week 14
      Research pairs submit a 5-minute video in which they present preliminary findings from their work and receive feedback from another research pair via the same rubric. (Students earn points for submitting a recording AND for providing feedback.)
   5. Week 15
      Research pairs make final adjustments to their poster and submit it, along with a copy of their research paper, for final evaluation by the instructor via the same rubric. (This is the final product, which will hopefully be higher in quality due to the scaffolded nature of the assignment, the sustained team effort, and two rounds of formative peer feedback.)

Magic School AI offers tools that can speed up the process of designing lesson plans, writing assignment instructions, creating rubrics to evaluate student work, remixing tests, and more. A free version is available.