What is Generative AI and how does it impact Teaching & Learning?
Generative Artificial Intelligence (AI) is a branch of computer systems that can simulate human conversation. AI can answer questions, provide information, produce code, and engage in open-ended discussion.
AI is quickly becoming ubiquitous in our world. It is available to students whether we actively choose to address or include it in our courses or not. Similar to Google Search, AI can answer questions upon request. It has the potential to save time, foster innovation, promote critical thinking, and offer feedback. Therefore, it’s essential to recognize its presence, establish clear guidelines, and provide direction for effective use. Some popular generative AI tools include Microsoft Copilot (UW supported), ChatGPT, and Google Gemini.
Guidance from the College of Engineering
The College of Engineering leadership supports the integration of AI in courses and encourages instructors to discuss suitable AI usage with students, particularly in programming or writing-intensive courses.
The Learning Design and Technologies (LDT) team is committed to promoting the ethical use of AI, emphasizing privacy protection and academic integrity. We continuously evaluate AI technologies to align them with ethical principles and industry best practices, and we encourage the use of AI to enhance learning environments when appropriate.
Using Microsoft Copilot for Teaching and Learning at UW
Microsoft Copilot is an AI-powered assistant available to instructors, staff, and students at UW. This tool can help you write and summarize text, write code, generate images, edit presentations, and more. Unlike other AI tools, Copilot at UW will not train on the information you input. However, it’s important to use the tool responsibly and avoid inputting sensitive data, trade secrets, etc. For more details, please see the Learning Design and Technologies Use of AI statement in the next section as well as the Generative AI @ UW–Madison: Use & Policies page.
To access copilot, visit the UW microsoft co-pilot page linked here. Log in using your UW credentials to gain full access and start using the tool.
It’s important to evaluate your learning environment to determine how copilot can best serve you and your students. We encourage all instructors interested in the tool to consult with a member of the LDT team who can assist with determining how you might use it. Here are some key considerations to keep in mind when getting started with copilot:
- Avoid inputting sensitive or confidential information, such as student records/information, personal data, or trade secrets.
- Use copilot to support, not replace, original academic work.
- Always review and verify the content generated by copilot. While it’s a powerful tool, it may not always produce accurate results.
- Stay updated on any new features or changes to co-pilot by regularly checking in with the Learning Design and Technologies team and UW’s AI and technology support pages.
- Ensure that the use of copilot does not create barriers for students with disabilities. Make use of accessibility features and consider alternative methods for students who may need them.
- Provide clear instructions and support for all students either in your syllabus or unit overview to ensure they can effectively use copilot.
7 Strategies for Teaching with Copilot
This is an accordion element with a series of buttons that open and close related content panels.
1. Explore and evaluate how you and learners might use copilot
Consider the following:
- How might copilot help learners achieve the course goals and learning outcomes?
-
Copilot can assist learners by providing support and resources, helping them understand difficult concepts, and offering instant feedback on assignments or other course materials. It can generate practice questions, summarize key points from lectures, and suggest additional readings, all of which can enhance students’ learning experience.
-
- How might copilot help you develop course materials, assessments, learning objectives, etc.?
-
Copilot can streamline the creation of course materials by drafting lecture notes, creating outlines, and generating presentation slides. It can help develop assessments by suggesting question types, creating quizzes, and providing grading rubrics. Additionally, copilot can suggest clear and measurable learning objectives that align with course goals.
-
- How can copilot be used to enhance the accessibility of course materials for diverse learners
-
Copilot can enhance accessibility by generating content in various formats, such as text, audio, and visual aids, to help align to different learning preferences. It can assist in creating transcripts for video lectures, simplifying complex texts, and translating materials into multiple languages.
-
2. Address copilot in your Syllabus/Course Overview
Please see the sample language provided by the LDT team below:
The use of generative artificial intelligence (copilot) in this course is allowed at the discretion of the instructor & should be used with the intent to enhance learning experiences. However, the use of AI raises important ethical considerations that must be addressed in order to ensure that it is used in a fair, responsible, and effective manner. Knowing this, learners should be transparent about the use of AI in this course, cite AI when it is used in course materials, be aware of potential biases provided and/or incorrect responses, ensure the privacy and security of data, and continuously evaluate its effectiveness. AI should not be used to generate content that is harmful or offensive to others. Currently employed engineers and other working professionals should identify any policies their companies may have that prohibits providing AI with company-specific information (business plans, trade secrets, sensitive or restricted data, etc.). Additionally, AI is to be used as a supplemental resource and should not replace independent study and critical thinking. By using AI effectively & responsibly, you acknowledge the importance of maintaining high ethical standards & academic integrity.
3. Discuss copilot as a class
Prompts to discuss AI as a Class:
• Introduce the concept of AI and explain what it is.
• Discuss the accuracy and limitations of AI, as well as any biases that may be present.
• Discuss potential ethical and privacy concerns around AI, such as the risk of generating harmful or misleading content and the potential for data privacy violations.
• Encourage students to think about the broader implications of artificial intelligence in general: How might these technologies change the way people live and work in the future, and what steps can we take to ensure that they are used ethically?
4. Inform Students How to Cite AI
Considerations for Citing Generative AI: https://researchguides.library.wisc.edu/c.php?g=1334141&p=9825297
APA Style
In-text citation example:
(OpenAI, 2023; see Appendix A for the full transcript)
Reference section example:
OpenAI. (2023). ChatGPT (Mar 14 version) [Large language model]. https://chat.openai.com/chat
5. Design Assessments with AI in mind
A valuable opportunity involves encouraging learners to critically evaluate AI–generated outputs.
For example, ask learners to complete an assignment using AI and evaluate the result for accuracy, or have them refine and augment the AI–produced content to create an enhanced final product.
Consider these other applications:
• Using ChatGPT for Assignments | Tips & Examples
• Generative design: an innovative approach to design that involves using algorithms and artificial intelligence (AI) to generate and evaluate multiple design options.
• Code generation, review, analysis, and debugging: generate code snippets to reduce manual coding efforts; analyze code to identify potential errors, vulnerabilities, and inefficiencies; identify the root cause of errors or unexpected behavior in code more efficiently.
• Simulation and modeling: develop simulations and models of complex engineering systems or processes.
• Data analysis: analyze large data sets and extract meaningful insights.
• Project management: Task and resource scheduling; Risk Management; Decision–making.
• Technical writing: improve the clarity, coherence, and accuracy of technical writing.
• Collaborative writing: construct or expand upon ideas to improve on a collaborative writing effort
6. Privacy and Academic Integrity
For learners who have privacy concerns and don’t want to create an AI account with their name and/or email with an AI tool, consider alternative strategies and tools outside of AI: Discussion forums; Peer–to–peer collaboration; Kaltura, and YouTube: There are many educational Kaltura and YouTube channels that offer tutorials and lectures on various subjects to explore.
For instructors who may have concerns about academic integrity, consider the following:
• Encourage originality by designing assessments that require students to draw on their
own experiences, thoughts, and observations.
• Focus on problem–solving and higher–order thinking skills in assignments, prompting
students to synthesize information, make connections, and critically evaluate ideas.
• Ask learners to incorporate course texts and resources into their responses.
• Explore collaborative assignments that rely less on AI.
• Break writing assignments into smaller assessments, incorporate peer review.
7. Guidelines for Effective Prompting with Copilot
- Specify Desired Output
- Clearly articulate what type of response you’re looking for in your prompt. If you want a specific format, length, or style of response, specify that in the prompt itself. The AI relies on your instruction to generate the most appropriate response.
- Avoid Ambiguity
- Ambiguity can lead to a wide range of responses from the AI, not all of which might be helpful. Be as explicit as possible in your language to avoid misinterpretations, and remember to ask direct questions if you need specific information.
- Use Contextual Information
- Context is key in prompt creation for AI. Include necessary background info for AI to grasp the situation. However, remember, models like ChatGPT don’t access personal data unless shared during the conversation.
- Refine Prompts Over Time
- If the AI’s response isn’t what you were hoping for, try rephrasing or adding more details to your prompt. Consider guiding the conversation and provide further instructions based on the outputs you receive.
- Balance Prompt Details
- Provide enough information in your prompt for meaningful responses, but avoid unnecessary details. Aim for a balance of context and clear instructions.
Leveraging AI for Accessibility: Enhancements & Challenges
Personalized Learning
AI can help create a more personalized learning experience, adapting to each student's learning style. This can be beneficial for students with disabilities, who may require a more tailored approach.
Technological Accessibility
Not all AI technologies are accessible to all students. For example, certain AI-powered tools may not work with assistive technologies like screen readers. It's important to choose AI tools that follow accessibility standards.
Assistive Technologies
AI is helping with assistive technology. For example, AI-powered speech-to-text and text-to-speech technologies can assist students with hearing or visual impairments.
Digital Divide
Not all students have the necessary technology or internet access to use AI-powered tools, which can create or widen gaps in learning. It is important for instructors to recognize this in their learning environments and create a course of action.
Learning Analytics
AI can analyze student data to identify areas of difficulty and provide timely feedback, helping students who may struggle in certain areas.
Data Privacy
AI often relies on collecting and analyzing large amounts of data. Ensuring this is done in a way that respects privacy rights and is accessible to students with various abilities can be a challenge.