AI Learning Tools: Instruction Implications

• There are productive, ethical uses of AI tools. In fact, many industries are already using AI and will expect new hires to understand how to use AI effectively.
• We all have a role to play in getting students to understand when AI use is appropriate and when it is not.
• We need to stress the benefits of students doing their own work:
  • The writing process helps students synthesize information and learn how to apply it to their own lives and careers.
  • The writing process develops critical thinking skills as well as communication skills.
• Likely outcomes if students do not generate independent work:
  • Unable to speak well about the discipline and their knowledge/skills during an interview
  • Will struggle to perform on the job – ChatGPT, online test banks, word spinners, ghost writers will not be there to help them be efficient and productive in industry. It will not help them write work reports, communicate intelligently with peers and supervisors, or respond to urgent, in-the-moment situations that call for knowledge and critical thinking skills.
• Importance of programs/departments talking about this and coming to a consensus on approach so there is consistency within a program/department.

Although many faculty are understandably concerned the academic integrity implications of students using tools like ChatGPT (Marche, 2022), generative AI can also be used to enhance and support teaching. Here are some of the ways that you can incorporate AI tools into your teaching:

• Generate responses to common student questions or emails.
• Create test questions or test question answer options.
• Draft lesson plans or assignment instructions.
• Generate feedback comments to create a comment bank for assignments.
• Create examples or samples for students to compare their own work to or analyze.
• Walk students through how to use ChatGPT like a sophisticated version of Grammarly to revise and edit their writing.
• Demonstrate how to use ChatGPT as a basic tutoring tool to help students get explanations for confusing concepts.
• Use ChatGPT to give students’ real-time feedback on their writing in language learning classes.
• Summarize qualitative student feedback from OMETs.

• Break longer assignments into parts. Make sure that all parts are cohesive in the end.
• Require students to show or explain their work:
  • Oral presentation or instructor-student conference to discuss the written paper: providing a summary of what they learned, which resources were most helpful and why, and why they organized the paper the way they did.
  • Provide drafts along the way, starting with outline. Provide feedback so that there are specific areas for students to redevelop. Will require faculty to compare to ensure that there is solid connection between drafts and development of the same ideas.
  • Require students to provide their notes.
• Turn smaller written assignments into classroom assignments.
• Require more hand-written, in-class responses to prompts to assess learning.
• Incorporate reflective elements that require the student to apply didactic elements to specific experiences in lab and class.
• Require sources for all online discussions – the system doesn’t handle this well, so it may be a good way to catch AI-generated content.
• Written assignments that include interviews with local professionals that can be verified.
• Written assignments that have a local connection.
• Written assignments that require the students to draw upon classroom discussion.
• Written assignments based on very recent (past 6-12 months) events.
• Have students start a writing assignment in class; they may be less likely to start over if they already have a thread started.
• Enter your prompt/assignment description into ChatGPT and retain the results. If the same parameters are used by the students, the output will be very similar in structure, flow.
• Try creating assignments that use ChatGPT to encourage critical thinking. For example, Have the students generate a paper using ChatGPT and then require them to analyze the output, identifying the holes, the inaccuracies, logic problems, etc.

Tools for detecting AI-generated text are still developing. You can use these tools as a starting point, but be aware that results may not be accurate as the tool continues to develop.