Our Winter AI Institute for Teachers
For the second time I had the pleasure of designing an AI Institute for Teachers at the University of Mississippi. You can read more about the first AI Institute we designed and ran last summer that was profiled a few weeks ago in the Washington Post. My wonderful chair, Stephen Monroe, and the Center for Excellence in Teaching and Learning and Academic Innovations Group co-sponsored the event. And what an event it was! Three days of activities for helped nearly 100 faculty across the university develop their AI literacy. Josh Eyler and Emily Pitts Donahoe held panels on alternative grading and crafting syllabus policy language for our generative AI era. You can access the full curriculum below—I’ve released all the materials under a CC-BY license so that they can be reused and remixed. Having hosted this AI literacy institute twice, I've developed an effective approach for offering faculty development opportunities on a rapidly evolving technology like generative AI.
How we ultimately define AI literacy in academic contexts may be far more fluid and situational depending on the stakeholder’s comfort level and knowledge. To be sure, one person’s AI literacy may be far more advanced than another, but one of the core purposes of this training institute was to create a sort of baseline—here’s what this technology does; what we know about how generative AI is made; how people are using it in education and the world; what ethical challenges this poses to all stakeholders. If you’d like to take a deeper dive into this frame of mind, please check out how Maha Bali outlined a similar approach in What I Mean When I Say Critical AI Literacy.
Where Implication Meets Application
Part of my job with our Academic Innovation Group is to consider the implication of new technology on teaching and learning. This is juxtaposed against the more established position of faculty development in a traditional Teaching Center, one that focuses on the application of existing techniques, technology, pedagogy, research, evidence, etc. Of course, the problem is generative AI occupies both implication and application within teaching all at once. So how do you design something for both?
Hands-on Learning
Stakeholders need hands-on workshops with the technology beyond chatbot interfaces like ChatGPT to explore its pragmatic potential and ethical challenges. Simply listening to a lecture or a slide deck isn’t enough. People need the opportunity to experience these tools first-hand, and that takes a great deal of time and thoughtful planning to design experiences. Here’s how I designed the activities and what I would do again:
Learning Outcomes:
Giving attendees a brief overview of what they’re doing and why it matters is crucial framing for each activity. When I was building these interactive portions, I had to ask myself:
what skills do I want faculty to develop to enhance their AI literacy?
what implication does this tool or technique have for learning?
How can this activity be used to showcase an ethical challenge with AI?
What are the perils and potential of deploying this tool in the classroom?
Often, many of the learning outcomes were weaved into each activity. I think this overlap matters a great deal. Mixing a healthy dose of ethics, implication, and pragmatic application helped establish many of the core principles of AI literacy
Short and Interactive:
Most people can handle 15-25 minutes of a discussion or an activity before things begin to drag. With this in mind, talk to them first about stakes, before having them experience the tool/ technique, then get them to share and reflect on it through discussion. The major challenge here is balancing the material we’re covering in such a way that isn’t overwhelming for faculty. That’s something I need to improve upon, which is a good point to segue to some challenges.
Areas to Improve
Yes, the exercises moved swiftly and having to present, circulate, and discuss was a major challenge. I move to one group and answer questions, then get pulled to the other side of the room to answer technical challenges or address and individual faculty members' concerns. However, the main challenge was onboarding for each activity. This is certainly something I need to revisit.
We Needed an Onboarding Session:
There were too many apps that faculty needed to sign up for, and too many times where logins asked attendees to use their private Google accounts. Many were uncomfortable for data privacy concerns to share such information. Next time, I’ll create a pre-session checklist, asking faculty to create secondary Google accounts so that they can sign in to many of these apps without concern. I imagine spending 20 minutes with an onboarding session to go over this material will be quite helpful.
AI Institute for Teachers Curriculum
Over three days, dozens of faculty members from across schools and disciplines came together to explore how generative AI was impacting their teaching. For the second time, the University of Mississippi’s Department of Writing and Rhetoric hosted an AI Institute for Teachers for its faculty and open it more broadly to the university. Three days of highly interactive workshops gave educators opportunities to use the tools hands-on to generate text, audio, and video, and even create their virtual avatar. With each activity came a set of questions to critically center these new features—how will this technology or technique impact my students and my teaching?
Day One: AI Literacy Basics: Slides, Activities, and Assignments
Introduction to Generative AI:
How neural networks and Large Language Models (LLMs) function.
Activity: Google’s Quick, Draw
The complexity involved in crafting effective prompts for AI.
Activity: Say What You See
Critical insights into the limitations and potential biases of these models.
Strategies for assignments that help students develop AI literacy, including analyzing and producing machine-made stories, and writing assignments that critically engage with AI capabilities.
Ethical Challenges Posed by Generative AI:
Understanding the broader scope of generative AI beyond text generation, including its ethical, legal, and societal implications. PPT of Ethical Challenges
Recognizing the potential for AI to be used in spreading propaganda and misinformation, and its impact on credibility.
Activity: The Bad News Game
Activities to develop AI literacy for images and broader ethical challenges posed by generative AI.
Activity: Which Face is Real?
Activity: Ethical Challenges
AI Detection:
The challenges and limitations of using AI to detect AI-generated content. PPT AI Detection: An Arms Race
Why AI detection is ineffective in academic contexts.
Activity: Real or Fake Text
Activity: Test an AI Detector
The importance of scaffolding assignments, building relationships with students, and focusing on learning not grades, as potential methods of mitigating AI’s impact on teaching.
Day Two: Applying AI Assistance: Slides, Activities, and Assignments
Language Models and Prompt Engineering:
An overview of different available LLMs and AI-powered applications approved by the university. So Many Language Models
Introduction to basic prompt engineering techniques: Introduction to Prompt Engineering PPT
Activities focused on creating personalized prompts and understanding the nuances of AI responses.
Activity: Creating Your Own Persona
Activity: Providing a Reference Text
AI Writing and Reading Assistants:
The impact of generative AI on writing and reading practices in academia.
Ethical guidelines for AI usage in academic settings.
Exploration of specific AI interfaces like Lex for writing assistance and ExplainPaper and SciSpace for reading assistance.
Introduction to Lex, an AI-Powered Writing Assistant
Activity: Explore use cases for Lex 20 minutes
Introduction to ExplainPaper and SciSpace, AI-powered Reading Assistants
AI Research and Speech Recognition:
The use of AI-powered research tools in academic contexts.
Understanding AI-powered speech recognition and its potential to support lectures and assist students with disabilities.
Explore the ethical challenges of open models by red-teaming a models for safety features
Activities involving AI research tools like Perplexity and discussions on ethical considerations in using AI in the classroom.
Introduction to Perplexity, an AI-powered research assistant
Red Teaming an Open Model:
Research Assignments:
Perplexity: Research Question Assignment
Perplexity: Stress Testing Claims Assignment
How Speech Recognition Can Support Students
Assignment:AI in the Classroom: An Ethical Debate
Day Three: Advanced AI Frontiers: Slides, Activities, and Assignments
Advanced Applications of Generative AI:
Exploring generative AI capabilities in creating images, videos, and voice generation: The Changing Landscape of Generative AI
Understanding how to effectively prompt AI for various use cases.
Exploring Video and Audio Generation
Activity: Create your own video story
Activity: Explore a Music Generator
Retrieval Augmented Generation (RAG):
Understanding why RAG is effective for certain tasks and its limitations.
The challenges of working with current transformer models.
Exploring advanced AI tools like Google’s NotebookLM and Anthropic’s Claude 2.
Activity: Explore Google’s NotebookLM
Activity:Explore Anthropic’s Claude 2
Building AI Chatbots:
Learning the basics of programming a system using natural language.
Understanding the potential and challenges of AI agents in education: PPT How to Build your own Chatbot
Ethical guidelines for using AI in educational settings.
Practical activity training a chatbot using Poe.
Connect to Learn More
I really hope the AI Institute for Teachers provides a robust model for developing AI literacy across disciplines that others can use. Please reach out if you’d like help in developing your own or would like me to visit your institution. I’m always happy to connect!
My email: mwatkins@olemiss.edu
The Path Forward
Focusing on pragmatic application, ethical implication, and critical analysis, attendees gained relevant skills to evaluate how generative AI may impact their teaching. The hands-on, interactive format allowed for meaningful discussion and reflection on these rapidly evolving technologies.
As AI continues advancing, such professional development opportunities will only grow in their importance. We must equip faculty to harness the potential of AI while safeguarding quality education. This means putting AI skeptics and AI advocates in the same room together (in-person or virtually) so that we don’t silo ourselves and stunt this evolving discourse. Please use the materials provided to establish a springboard for your own customized workshops and tailor them to your need.
Ultimately, AI literacy will be a key concept for educators in 2024 and beyond. We have an opportunity to collaboratively shape how AI is used in education to promote creativity, critical thinking, and human dignity, but very little time to do it. The future of a just and democratic society will depend on how well we prepare faculty and students today and AI literacy is just the latest competency that needs to be addressed.
Marc, this is a spectacular resource. Thank you for sharing it.