Pages

Using ChatGPT for Syllabus Refresh

Syllabus Refresh – Prompt Engineering for ChatGPT

Notes from OTL’s session for the AI in Action series.

Kim DeBacco and Kate Schaller worked with faculty to generate and regenerate a course syllabus. Together they investigated prompt engineering to iteratively refine their choices.

Session materials:

Ask Chat GPT to create a syllabus

Initial prompt:

Design a syllabus for an introductory, undergraduate-level university course about "The Psychology of Aging".

Sample refining prompt:

Regenerate this syllabus to include the following syllabus sections:

Course title: “The Psychology of Aging” (PSY136) 
Instructor Information: Kim DeBacco, PhD; kdebacco@online.ucla.edu
TA Information: Qiwen Moore, David Christomakis. 
Contact Information: tba
Course Prerequisites: PSY124
Course Description:
Learning Outcomes for this Course:
Course Materials:
Technical Requirements (Bruin Learn, browsers etc.)
How to Succeed in this Course (Expectations for Students, Study Advice & Tips)
Creating an Inclusive Classroom Community (Instructor, TA, and Community Expectations)
Course Schedule (Dates & Topics, Readings)
Predictable Weekly Pattern (for Students)
How Your Learning Will Be Assessed (Grading Policy)
Information about the Course Assignments
Link to the Quarter dates in the UCLA Academic calendar 
The class dates and times: Tuesday 10am - 12noon; Thursdays 10am - 12 noon
Link to UCLA resources for Student Well-being
Link to UCLA Policy on Academic Integrity

Model the prompt by inputting two or three past syllabi:

Integrate information, readings and topics from this syllabus. Do not include the dates.

NOTE: The example syllabi were freely provided, and previously downloaded from merlot.org

Additional Refining Steps

Provide a list of readings with links; link to a textbook, etc.

Add your lists of readings – Required/Essential, Recommended, etc.

Integrate these readings into the syllabus schedule:

Park, D.C., K. Warner Schaie, K. Schaie, S. L. Willis, S. Willis. Handbook of the Psychology of Aging. 7th edition.

Provide a focused, scripted course description, to narrow and specify content.

Regenerate the syllabus to reflect this description: [your course description goes here]

Read closely and correct!

  • Add your emphases, your voice and values, special sections, and more.
  • Consider sharing your syllabus with a colleague for feedback.

Solicit student feedback

After the quarter is over, ask (some of) your students to give you annotated feedback on your syllabus. Ask them to add comments all over it:

  • what worked well;
  • what advice would they give future students;
  • what would they change;
  • which readings and media made sense;
  • which readings were tough but worthwhile;

and so on…

See also the original session document.

ChatGPT-Generated Quiz Instructions

Advice from ChatGPT on how to best use ChatGPT to create quiz questions

Instructions Generated by AI

Here are some instructions for college instructors to use ChatGPT to author low stakes quizzes. By following these steps, you can use ChatGPT to create low-stakes quizzes that assess your students’ knowledge and understanding of course concepts.

Identify the Topic

Decide on the topic for the quiz. This could be a specific concept or a broad subject area.

Determine the Learning Objectives

Identify the learning objectives you want your students to achieve through the quiz. These objectives should be specific, measurable, achievable, relevant, and time-bound (SMART).

Choose the Quiz Type

Determine the type of quiz you want to create. This could be a multiple-choice, true/false, or short answer quiz.

Provide ChatGPT with the Question

Use ChatGPT to generate questions related to the learning objectives and the chosen quiz type. You can provide ChatGPT with a prompt, such as “Please generate five multiple-choice questions about the topic of X,” and it will provide you with questions to choose from.

Review and Edit Questions

Review the questions generated by ChatGPT and edit them as needed. Ensure that the questions are clear, concise, and aligned with the learning objectives.

Create Answer Choices

If you’re creating a multiple-choice quiz, provide ChatGPT with answer choices for each question. Ensure that the answer choices are plausible and that one option is clearly the correct answer.

Finalize the Quiz

Once you’ve reviewed and edited the questions and answer choices, finalize the quiz by creating a clear and concise set of instructions for your students, specifying the time limit, grading criteria, and any other relevant information.

Administer the Quiz

Administer the quiz to your students through your preferred learning management system (LMS) or other means of distribution.

Grade the Quiz

Once the quiz is completed, grade the responses according to the grading criteria you established.

Provide Feedback

Provide feedback to your students on their performance on the quiz, including areas where they did well and areas where they could improve.

Suggestions for getting better results when generating quiz questions using ChatGPT

Be Specific with Your Prompts

The more specific your prompts are, the better the results will be. Instead of asking ChatGPT to generate questions about a broad topic, try to provide a specific question or scenario to prompt the AI. This will help ChatGPT understand what type of questions you’re looking for and generate more relevant and focused results.

Provide Additional Context

Providing additional context to ChatGPT can help it generate more relevant and accurate questions. For example, if you’re asking ChatGPT to generate questions about a historical figure, provide some additional context about their life, achievements, and legacy.

Use Proper Grammar and Syntax

ChatGPT works best with clear and grammatically correct input. If you use improper grammar or sentence structure in your prompts, ChatGPT may generate questions that are difficult to understand or nonsensical.

Choose the Right Quiz Type

Different quiz types require different types of prompts. For example, if you’re creating a multiple-choice quiz, provide ChatGPT with answer choices for each question. If you’re creating a short-answer quiz, provide ChatGPT with examples of possible answers to the question.

Review and Edit Generated Questions

ChatGPT is not perfect and may generate questions that are unclear or irrelevant. It’s important to review and edit the generated questions to ensure that they align with your learning objectives and are appropriate for your students.

Using ChatGPT to Write Quiz Questions

Using ChatGPT to Write Quiz Questions

Notes from OTL’s session for the AI in Action series.

Mark Kaysar and Agustin Ríos used AI to create quiz questions with different question types. They also prompted ChatGPT to properly format a quiz for uploading into Bruin Learn.

Can ChatGPT help faculty create and format low stakes assessments for using in Canvas?

Demo Question Generation

  • Question Types
  • Multiple Choice
  • True/False
  • Fill in the blank
  • Short Answer
  • Short Essay
  • Adding feedback to answers

Quiz Formatting

When working in ChatGPT, if you include the format instructions shown below you will receive questions that are ready for upload to Canvas. ChatGPT will usually respond in the format you tell it to use, but if it does not, resubmit and ask it to reformat.

Please create all quiz questions using the following format. Each choice needs to start with a lower case alphabet, 
a, b, c, d, etc. with a close parenthesis. The correct choice is designated with an asterisk.

1. What is 2+3?
a) 6
b) 1
*c) 5
d) 10

Sample ChatGPT Prompts

Provide a topic

Can you write some multiple choice questions about [Insert topic]?

Provide a link to a reading

Please write multiple choice questions to evaluate this content.
[paste content]

1.2 The Weakness of Early Patent Systems - Introduction to Intellectual Property | OpenStax

Provide a link to a Canvas page

Please write multiple choice questions to evaluate this content.

Can you provide feedback on why the incorrect items are incorrect?
[paste content]

https://canvas.ucdavis.edu/courses/34528/pages/being-present-in-your-online-course

Provide a desired outcome

Please write multiple choice questions that evaluate this outcome [paste outcome]

Analyze how the costume designer’s interpretation of the screenplay is affected by the 
tone of the politics of the era and the pressures from the studio

Provide a video script or video captions

Please write multiple choice questions to evaluate this content.

[paste content]

Provide a set of answers

Please write multiple choice questions that have the following as answers.

[paste answers]

Provide similar questions to use as a model

Please write multiple choice questions that evaluate the same content as this question.

[paste question]

Uploading to Canvas

Before you can upload the quiz into Canvas, you need to save the questions in the QTI file type.

Step one

Step two

This will create a quiz with your new questions.

See also the original session document.

Using ChatGPT in Conversations with IDs

Using ChatGPT to Lead In-depth Conversations with Instructional Designers

Notes from OTL’s session for the AI in Action series.

Sirui Wang and Brittany Goodwell explored possible ways to use ChatGPT to start thinking about different teaching and learning challenges using two scenarios to begin instructional design conversations between faculty and instructional designers. Both scenarios included questions and requests often asked by faculty. There were opportunities to interact with ChatGPT and instructional designers.

How can ChatGPT be involved in the course design process?

  • ChatGPT integrates related information and output in a well-structured way. The advantage of using this generative AI enhances the consultation process with higher efficiency.
  • The conservation style allows the information to be built upon one another, forming a useful dialogue designers can share later.
  • Using ChatGPT in the initiative process of an instructional design consultation helps focus on what faculty needs and narrow the topics, making the discussion with instructional designers more efficient and allowing in-depth discussion.

However, what ChatGPT provides is usually very surface and general, and instructional designers must step in and continue the in-depth conversation.

Instructional Designers’ Roles in the AI Age

  • In-depth conversation
  • Connect with the real course design
  • Ensure equity and inclusion
  • Create personalized learning experiences that adapt to the needs of individual learners
  • Incorporate gamification and simulations to make learning more interactive and engaging
  • Share experience

Scenario 1 – Identify the Teaching and Learning Challenges

ChatGPT Screenshot 1: Engaging students more effectively

How to use ChatGPT to begin the conversation?

  • Start with the basic question/challenge
  • Go through the options, and see if there is anything that interests you
  • Ask for or think of more specific questions, such as #5, why frequent feedback will engage students more efficiently, how to provide frequent feedback, what types of feedback can be provided, etc.

But ChatGPT does not have the accurate context for what it outputs, so a continued conversation with instructional designers are strongly encouraged.

How to continue the topic with an Instructional Designer?

  • Discuss and continue the specific questions: Why feedback loop is important, especially online? What feedback loops can and will include?
  • Clarify the focus: what could be a good fit for your class? How do involve the three parties in a class: instructor, TAs, and students?
  • The deeper conversations: how does it align with other class pieces?
  • Access to other successful course design experiences from different courses

Scenario 2 – Educational Technology Recommendations

Screenshot 2: ChatGPT advice on good class technologies
Screenshot 3: Refining ChatGPT session

How to continue the topic with an Instructional Designer?

  1. Review and refine the recommendations: The faculty member could work with the instructional designer to review the recommendations generated by ChatGPT and refine them based on their specific course needs. The instructional designer could help the faculty member evaluate the options based on factors such as ease of use, accessibility, cost, and pedagogical value.
  2. Identify and address potential challenges: The faculty member and instructional designer could discuss potential challenges that may arise from implementing the selected educational technology in their course. This could include issues related to technology integration, student engagement, and support for students who may have varying levels of technological proficiency.
  3. Explore implementation strategies: The faculty member and instructional designer could work together to explore implementation strategies for the educational technology recommendations.
  4. Determine next steps: The faculty member and instructional designer could determine next steps for moving forward with the implementation of the selected educational technology. This could include identifying any additional support or resources needed, setting timelines for implementation, and designing appropriate activities.

See also the original session document.

AI in Action – Course Design Opportunities with AI

Course Design Opportunities with AI

On May 16, 2023, as part of the campus series, “AI in Action: Exploring AI’s Potential in Teaching and Learning,” OTL instructional designers worked with over 30 participants across three breakout rooms to explore ways that AI can be used to enhance teaching and learning.

  • Using ChatGPT to Lead In-depth Conversations with Instructional Designers: Sirui Wang and Brittany Goodwell explored possible ways to use ChatGPT to start thinking about different teaching and learning challenges using two scenarios to begin instructional design conversations between faculty and instructional designers. Both scenarios included questions and requests often asked by faculty. There were opportunities to interact with ChatGPT and instructional designers.
  • Using ChatGPT to Write Quiz Questions: Mark Kaysar and Agustin Ríos used AI to create quiz questions with different question types. They also prompted ChatGPT to properly format a quiz for uploading into Bruin Learn.
  • Syllabus Refresh Using Prompt Engineering in ChatGPT: Kim DeBacco and Kate Schaller worked with faculty to generate and regenerate a course syllabus. Together they investigated prompt engineering to iteratively refine their choices.

AI in Action: Course Design Opportunities with AI

May 16 @ 10:00 am 11:00 am

OTL’s entry in the campuswide series, AI in Action: Exploring AI’s Potential in Teaching and Learning

Join us to explore some of the ways to leverage AI to find resources, design assignments, give students feedback and improve student writing.

Calendar invites containing a Zoom link will be sent to those who register for the May 16th workshop by completing this form.

Please email us at contact@online.ucla.edu if you have any questions.

Series: AI in Action: Exploring AI’s Potential in Teaching and Learning

May 12 @ 12:00 pm June 2 @ 12:00 pm

Seven events from seven campus units exploring how Artificial Intelligence can impact teaching and learning.

UCLA’s Center for the Advancement of Teaching (CAT), the Center for Education, Innovation, and Learning in the Sciences (CEILS), the Excellence in Pedagogy and Innovative Classrooms program (EPIC), Online Teaching and Learning (OTL), the Bruin Learn Center of Excellence (COE), the Writing Programs, and Humanities Technology (HumTech) have collaborated to bring this event to you.

The kick-off event for AI in Action will be held on Friday May 12th at 2pm PT, followed by a week of follow up workshops that dive deeper into potential uses of AI in teaching and learning. Events held from May 12-19 will take place remotely over Zoom.

Instructional Design Showcase – SCAND 60

May 4 @ 10:00 am 11:00 am

This is an online event  

Register to receive the Zoom link and calendar invitation

What to Expect in this Showcase

In this course showcase session, the instructor will share his experience designing a fully asynchronous online course on Nordic cinema based on earlier experiences teaching similar courses in person. The aims of the in-person and online courses remain the same: to offer undergraduates a broad introduction to a lesser-known cinematic tradition while also engaging them in critical thinking and writing. Topics to be covered include:

  • Employing various pedagogical strategies, such as podcasting, peer review, and blogging, to foster student engagement and build a vibrant learning community.
  • Encouraging students to have informal discussions through podcast activities to increase their engagement and interaction.
  • Effectively managing asynchronous schedules for students.
  • Utilizing Bruin Learn to facilitate a successful transition from in-person to asynchronous learning in the Introduction to Nordic Cinema course. 

Presenter Bio: 

Patrick Wen is a continuing lecturer in Scandinavian within the Department of European Languages and Transcultural Studies. He teaches courses on film history, modern literature, and also Norwegian language.

ID Showcase – SCAND 60 – Introduction to Nordic Cinema

Scandinavian Studies 60 – Introduction to Nordic Cinema

SCAND 60 course screenshot

In this course showcase, Prof. Patrick Wen will share his experience designing a fully asynchronous online course on Nordic cinema based on earlier experiences teaching similar courses in person. The aims of the in-person and online courses remain the same: to offer undergraduates a broad introduction to a lesser-known cinematic tradition while also engaging them in critical thinking and writing.

Topics to be covered include:

  • Employing various pedagogical strategies, such as podcasting, peer review, and blogging, to foster student engagement and build a vibrant learning community.
  • Encouraging students to have informal discussions through podcast activities to increase their engagement and interaction.
  • Effectively managing asynchronous student schedules.
  • Utilizing Bruin Learn to facilitate a successful transition from in-person to asynchronous learning in the Introduction to Nordic Cinema course.

Presenter Bio:

Prof. Patrick Wen is a continuing lecturer in Scandinavian within the Department of European Languages and Transcultural Studies. He teaches courses on film history, modern literature, and also Norwegian language.

ID Showcase – Engaging STEM Students with Perusall – CLUSTER M71 and SOC GEN M144

Engaging STEM Students with Perusall – CLUSTER M71 and SOC GEN M144

In this course showcase session, the instructor will compare two different Biology and Society courses that have integrated Perusall as one of the engagement solutions for a large general education course and an upper division elective in the Human Biology and Society major.

The presentation will include the following aspects:

  • How Perusall has been selected and integrated
  • Benefits for collaborative learning and critical thinking 
  • The challenges of using Perusall in interdisciplinary science courses

Presenter Bio: 

Prof. Michelle Rensel is an adjunct assistant professor in the Institute for Society and Genetics (ISG) and the coordinator of the freshman Biotechnology and Society GE Cluster course. Prior to joining the Institute, she completed a PhD and postdoctoral work in behavioral endocrinology. As full-time teaching faculty in ISG, she teaches a broad range of upper and lower division courses that bridge the life and social sciences, and regularly implements new instructional strategies to improve learning, belonging, and retention in the life sciences.

Presented: Wednesday, April 26, 2023

Selected Q&A from the Showcase


QuestionAnswer
What is the average enrollment in the cluster?Cap enrollment is 240 (usually reached); 6 TAs all year; sometimes a little dropoff in enrollment in winter but little from winter to spring.
If copying a course to a different quarter, will the Persuall set up be copied over too, such as the assignments list, instruction, etc.?Yes – you can select what you want to copy! (You could just copy over the library, or you could copy individual readings, or you can copy the whole thing).
How much do you/other faculty in the cluster annotate in Perusall, or comment on annotations?Prof. Rensel usually does not intervene. Sometimes students tag her and ask a question, and then she responds. But usually not! If she intervened, it would be in a smaller course to guide discussion. Students also benefit from feeling like Perusall is “their thing”.

There haven’t been issues so far with students saying anything horrible – they spend time talking about codes of conduct.

It seems that you have to manually build individual assignments in Perusall, is that correct?Yes and no. Once you add your resource you can create your annotation assignment for that resource. Please note that you can use your Perusall course scoring settings for all your assignments – you do not have to set your scoring settings for each individual assignment.
Do students see other people’s annotations before they post their first comment? Did you have any issues with repeated points or syntax?Students have noted plagiarism by other students. Students can choose to turn off seeing other annotations but they need to make that choice.

Oftentimes there are double annotations (could be from wifi issues or glitches) – there is a setting to automatically flag duplicates.

Is the grading that happens in Perusall automatic?Yes – you can set (and later change) the scoring settings and the annotations are auto-graded. You can set a formula in the grading scheme so students have multiple ways to reach full credit. You can also adjust students’ scores before releasing them to the gradebook.
How is “Reading to the end” determined?The reading metric gives credit based on how many pages the student has viewed. If the student only viewed half the number of pages, they would receive 50% of the score.

If students are commenting on a video or podcast, students will receive full credit upon opening the assignment as there are no pages in a video or podcast assignment.

Does an instructor need to create the scoring columns? Is there a limitation of how many columns/criterion? Or is that fixed?The columns in the scoring criteria show up as fixed options – you set the percentages of what things are worth.
What is roughly the distribution of scores you might expect?The goal is that everyone gets 100% – they do the reading and they do their annotations! Occasionally some get flagged as plagiarism – re: Michelle, Perusall doesn’t do much to detect that but students themselves have detected it! It’s not the norm but it is a possibility.

There is continued discussion in the class about best practices for using Perusall, so students are aware of how to use it.

Is there a way to filter by section so TAs can read over before their section?You can set up groups by section – but it was manual! There is a way to manually import now. You can import Canvas/Bruin Learn groups to Perusall.

If students move groups, their initial comment threads move to the new group and disappear from the old one.

When students are working on an assignment, do they have the same view as you are showing here? The texts are all covered by annotations.No, they do not have the same view; they can see each other’s within their groups.

There is an option to see what comments display. “My Comments” is the menu item that includes the flexible view of how they can view the comments.

Did students engage more with reading using Perusall compared to how they did reading before the tool?Students felt positive about using the tool and provide meaningful annotations.
How do you connect what the students post with your lectures and interaction with the students?With a smaller course you can use their annotations as inspiration for discussions. Other folks using it in seminars use it to have “questions in back pocket” when discussion stalls!

It is harder to make the connection in a big lecture course with lecture materials.

Do you know if it would be possible for students to upload materials that would become part of the assignment?In the “Library” section there is a “Student upload folder” so that students can share files directly. You may still need to put these materials into an assignment.

This Google Doc contains the comments, questions and answers collected during the showcase.

Events

Series: AI in Action: Exploring AI’s Potential in Teaching and Learning

May 12 @ 12:00 pm June 2 @ 12:00 pm

Seven events from seven campus units exploring how Artificial Intelligence can impact teaching and learning.

UCLA’s Center for the Advancement of Teaching (CAT), the Center for Education, Innovation, and Learning in the Sciences (CEILS), the Excellence in Pedagogy and Innovative Classrooms program (EPIC), Online Teaching and Learning (OTL), the Bruin Learn Center of Excellence (COE), the Writing Programs, and Humanities Technology (HumTech) have collaborated to bring this event to you.

The kick-off event for AI in Action will be held on Friday May 12th at 2pm PT, followed by a week of follow up workshops that dive deeper into potential uses of AI in teaching and learning. Events held from May 12-19 will take place remotely over Zoom.