Guidance for Artificial Intelligence (AI) and Large Language Models (LLMs) Assistance Technologies

The widespread accessibility of AI and LLM assistive technologies - including those that help students with  writing, coding, drawing, and solving equations/problems like ChatGPT and Google Bard - create new opportunities and formidable challenges for teaching and learning. These opportunities and challenges can sometimes be catalysts for changes to course assignments, instructional materials and policies (see UCSB's AI/LLM policy/guidance here). More information about AI for teaching and learning is also on the TALES AI 101 content. The guidance on this page is to help instructors teach students about the pitfalls and opportunities of AI within the context of their courses, as well as learn about how AI might be incorporated into their courses and work.

graphic of human coming out of a laptop screen tapping a sign for image prompt

Consider incorporating AI technologies into your course

  1. Create assignments where students use AI/LLMs as part of the writing/thinking activities. You can get inspiration by browsing 101 Creative Ideas to Use AI in Education and other resources.
  2. Stimulate discussions about problem-solving and writing processes, strategies, and ethics through class explorations of AI technology. Work with your students to generate diverse examples using AI, then compare and contrast them with student work, and examine their strengths and weaknesses collaboratively.
  3. Think about your course objectives. What are the cognitive tasks students need to perform without AI assistance? When could/should students rely on AI assistance to help them practice or find inspiration? Where can an AI facilitate a better outcome, creation, or product? Are new rubrics and assignment instructions needed?
  4. Explore ways that people in careers in your field (potentially) use AI. What kinds of assignments could help students learn those skills, as well as the pros and cons of using AI in their future workplace?
  5. Encourage student creativity and curiosity by leveraging AI technology to create initial prompts, classifications, topics, or questions for exploration.
  6. Challenge students to interrogate how AI technology can help them problem-solve, brainstorm, and compose pieces across various genres, styles, and perspectives.
  7. Invite students to utilize AI to generate text in specific genres in order to recognize and identify genre conventions and reflect upon the role of audience, purpose, and context in developing rhetorically effective prose.
  8. Encourage students to compare AI-generated text, code, analyses, etc. with human-generated versions to see how individual agency, voice, style, and ethos impact text.
  9. Examine the potential for and risks of integrating AI tools into the research process, given that LLMs can "hallucinate" and generate false facts, statements, or sources. Urge students to cross-check AI-generated information and develop critical appraisal skills to maintain the credibility and precision of their work.
  10. Examine and address potential biases and fairness concerns that may arise from AI technology, including the perpetuation of stereotypes or the exclusion of specific perspectives. Promote critical thinking and discussions to recognize and counteract biases in AI-generated content.

Key Approaches for Teaching with AI

Consider how you can set students up for success with the following approaches:

Communicate expectations for student use of AI in courses and/or other documents such as theses, dissertations, research articles, etc.

The use of AI writing technologies falls within the purview of the Student Conduct Code and the Student Guide to Academic Integrity. It states that “Materials (written or otherwise) submitted to fulfill academic requirements must represent a student’s own efforts unless otherwise permitted by an instructor.” Therefore, student use of AI-assistive technology is not allowed in coursework, on theses, dissertations, research articles, etc. unless specifically allowed by the instructor or supervisor. 
 
Whatever decision you make as an instructor:

  • Explain “why.” As AI tools become more integrated with commonly used programs (e.g. Google Docs), it becomes increasingly important for instructors to explain why and how AI tools should/should not be used. Considerations may include: the accuracy/credibility of AI generated work, potential bias of AI-generated results, developing students’ problem solving and critical thinking skills, individual voice/skills, etc.
  • Consider including a policy statement specifically about the use of AI tools (see examples in Sample Language for Syllabus Policies)

Use AI-Writing Assistance technologies for plagiarism detection, grading and feedback ethically. UCSB does not support the use of plagiarism detection software (e.g. Turnitin, ChatGPT Zero) for several reasons: 

  1. Anti-plagiarism software is highly fallible. LLMs are advancing at lightning speed with huge injections of capital. Procuring “anti” LLM software contributes to a virtual arms race, with detection software always one step behind what LLMs can produce.
  2. Submitting student work to anti-plagiarism software may violate students’ intellectual property rights.  When student work is uploaded into a AI-Writing/plagiarism detector database, the student may lose ownership of their work and the instructor/University unable to safeguard how it is shared and used in the electronic commons.
  3. Use of anti-plagiarism software can undermine the fundamental relationship of trust that must exist between learners and teachers. To move from “detection” of LLM use to “prevention,” instructors should consider how students can use LLMs as a tool to support their work and/or craft assignments and activities that cannot be produced by LLMs. While this approach may represent a shift in perspective or assignments, the Office of Teaching and Learning instructional consultants offer extensive support for instructors who would like to pursue this approach.
  4. Instructors and TAs should not use AI-assistive technology for grading and feedback unless the technology is supported by UCSB (e.g. use of GradeScope is permitted, as UCSB has a contract for its use and the technology has been vetted for FERPA compliance), for the reasons outlined above.

Avoid academic integrity issues stemming from student use of AI-assistive technologies through alterations to your assignments.

  1. Scaffold (writing) assignments, so that students are completing smaller pieces over time that will be incorporated into larger assignments. Also integrate opportunities for students to receive and incorporate feedback on those smaller tasks as they complete the larger assignment. Feedback can be from their peers and/or the instructor/TA.
  2. Add brief reflective writing tasks to assignments that ask students to analyze the choices that they made as they completed the assignment (give examples).
  3. Talk with students about the purpose of the assignments, how they will help them develop skills, how they relate to future courses or careers, and how they can succeed in the course.
  4. Work with students to use AI/LLMs in productive ways (see list above).
  5. Assign work that requires personal reflection or creative thinking. For example, ask students to describe and reflect on a personal experience related to the course material or have students create some media based on the content.
  6. As students complete (coding or problem sets) assignments, have them annotate their work by using commenting tools or recording their voice.

The Office of Student Conduct adjudicates academic and behavioral violations of the Student Conduct Code. If you suspect unauthorized use of AI technologies, submit an incident report. Be sure to include any samples of earlier/baseline student work to which the work in question can be compared.