Understanding AI Writing Tools and Their Uses for Teaching and Learning at UC Berkeley

How Does GenAI Affect Teaching and Learning at UC Berkeley?

Nationwide, educators have been engaged in thoughtful conversations about how important it is for students to engage with and learn how to use GenAI. Because GenAI applications vary so widely and have such vastly different use cases across fields and disciplines, the impacts of incorporating usage of GenAI in any one individual classroom context will also differ.

That said, GenAI may impact the work of teaching and learning in the following ways:

Instructors may want to address appropriate uses of GenAI tools in their class contexts. This may include adding language into a syllabus or for individual assignments to address explicitly how and when students may use GenAI for successful assignment completion.
Instructors may want to revise or rewrite course or assignment-level learning outcomes to mention explicit engagement with GenAI. Depending on the course context, usage of GenAI may fundamentally change the assignment goals and outcomes. It may benefit instructors to review and revise their course or assignment-level learning outcomes to anticipate whether students will engage with GenAI and, if so, what they will learn from engaging with GenAI. Alternatively, instructors may want to revise or review their learning outcomes to clarify what skills or competencies they hope their students will gain by not using GenAI, emphasizing what students should be able to do in their courses independent of GenAI usage.
Instructors may want to update course materials to include or refer to how GenAI may change practices and processes in their disciplines or fields. Certain course readings or materials may need to be updated to reflect changes in professional or disciplinary practices that have been affected by GenAI usage.
Instructors may want to include an explicit unit or lesson on conducting research in their courses to help students contextualize the use of GenAI as part of a larger research landscape. GenAI can be very effective at summarizing large swaths of information and generating output. However, GenAI output is not always accurate, and students may need to learn how to cross-check GenAI output with information from other sources, such as research databases and library-supported search engines. If research engagement is not already an explicit part of the curriculum, GenAI may motivate instructors to incorporate this engagement into their curriculum.

GenAI technology continues to evolve and it’s likely that advice about how, when, and when not to use these tools will continue to shift in kind. These are a few starting points that may help in conversations about student usage of this tool. Note that these ideas are intended to be educational and are not yet driven by any institutional policy. Review an overview of AI in Teaching and Learning at Berkeley as well as the UC Responsible AI Guidelines which outline the ethical use of AI.

In the remainder of this page, we outline a range of pedagogical strategies instructors can use to harness the power of the GenAI to further their learning goals.

This page will remain a work-in-progress and will be updated as use cases and engagement with Generative AI technology continues to evolve.

Last updated: June 26, 2024

Teaching Recommendations

Articulate a clear AI policy for your course.

Think critically about how, where, when, and why students might use GenAI in your class context. Pepperdine University’s Center for Teaching Excellence developed a tool to help you generate such an AI syllabus statement for your particular course.

Design assignment prompts and activities that refer to specific class materials and resources.

Output from GenAI tools tend to give highly generalized responses. Assignments and prompts that ask students to refer to specific course materials, such as individual lecture notes, readings, and unique case studies/problems, may mitigate inappropriate usage of GenAI.

Teach students how to cite ChatGPT and other GenAI tools.

There may be cases where using GenAI supports a student’s thinking and development on a particular assignment. If that’s the case, refer students to current citation guidelines:

Consider learning about “prompt engineering” to have a better understanding of what outputs are possible from GenAI tool usage.

GenAI’s output will vary depending on what kinds of prompts users input into these tools and systems. As such, it may behoove you to learn about or consider different approaches to prompting AI. For example, consider entering your own assignment prompts into different GenAI tools to see what kind of output you receive. Alternatively, consider some pedagogical approaches to prompt engineering that help you and your students recognize how AI can provide different perspectives on classroom activities and assignments while still supporting students’ active and critical thinking about GenAI’s output.

Fold in ChatGPT as an example of a tool that violates academic integrity when used inappropriately.

In a portion of the course syllabus dedicated to academic dishonesty, it is worth mentioning that if a student uses text generated from ChatGPT and passes it off as their own writing, without acknowledging or citing the influence of ChatGPT in their process, they are in violation of the university’s academic honor code. Lifting full sentences and paragraphs wholecloth, whether it’s from an encyclopedia, written article, or AI-generated text creation tool, is considered plagiarism. Providing concrete examples to students of what constitutes written plagiarism may help them make more informed choices about how and whether to use particular tools to support their writing.

Discuss the value of the writing process with students and help students see the value in writing as a skill/outcome/competency in your class context.

If learning to become a stronger writer is part of a course learning outcome or goal, weave that outcome intentionally in your Spring course. Help students understand the value or benefits they gain from writing in your class. For example, incentivize or give credit for students to turn in outlines, drafts, pre-writing, and other kinds of notes so that they demonstrate their thinking about a topic prior to the creation of a formal written piece. Scaffolding in pre-writing, drafting, or outlining will help faculty more easily see the evolution of a student’s thought before they engage with the final written product.

Consider giving students options for the media/mode of their assignments if possible.

If developing writing skills isn’t a specific need in your class context, consider whether it may be fruitful to give students options about how they may demonstrate expertise or understanding of course content. For example, provide options for students to complete an assignment as an audio memo, a poster, a slidedeck, or some other kind of creative project. Then, ask students to explain why they chose a particular medium for submitting evidence of their learning. That way, students can make intentional choices about how they’re choosing to express their ideas in your course context

Suggested Writing Prompts and Activities

Create an assignment where students analyze and critique what ChatGPT generates.

If you’re interested in having students actively engage with ChatGPT, invite them to input responses into ChatGPT and analyze what they notice about its output. What does ChatGPT do well in its response? What do they see as the limitations of the response? What are they noticing about the tone, style, and engagement with core ideas from the class you’re teaching? The more that you help students guide their own critique and engagement with ChatGPT’s output, the more thoughtful and informed students can be about what ChatGPT is capable of doing.

Design essay and exam prompts that require close discussion or analysis of the materials used for your class, including images, video, and other media.

Current users of ChatGPT have found that ChatGPT struggles most to generate text that incorporates analysis of cited materials or artifacts, such as images, video, and media. As such, if you are designing writing assignments that include “close reading” of particular texts or media that are relevant to your class context, it is unlikely that ChatGPT will be capable of producing very meaningful work.

Design prompts that require students to work with and incorporate multiple, cited sources in their writing.

AI-generated answers do well at providing expositions and facts (even if the “facts” are not all accurate!). However, when asked to relate one concept to another, ChatGPT presents an exposition of one concept followed by the other, making only minimal and superficial connections between them. You can strengthen this kind of assignment by specifying a need for students to engage with multiple sources, asking them to cite those sources and explain how they are connected around shared themes, arguments, and ideas in your course.

Create essay and exam assignments that require students to devote a significant amount of time and space to describing and analyzing a specific example, object, or case.

ChatGPT does an adequate job of summary, but is incapable of close reading and of description based on sensory perception. Any assignment that includes a clear and specific requirement to discuss concrete examples that cite sensory experiences will require students to do much of the work themselves.

Ask students to draft different GenAI prompts and examine the output.

Invite students to play with different prompts for GenAI tools and then ask them to analyze the differing outputs from those prompts. Invite them to consider how tweaks or adjustments to prompts may impact the output they see, and why that differing output may matter for different audiences and purposes in their writing.