By Nicole Corbin, Instructional Designer

The Emerging Pedagogies Summit is an annual event hosted by the Learning Innovation and Lifetime Education (LILE) group at Duke University, and I, Nicole Corbin, instructional designer at NSF Unidata, had the pleasure of attending.

This year’s event was packed with thoughtfully curated topics relevant to the NSF Unidata higher education community, including AI and workforce development.

—Emerging Pedagogies Summit 2024

AI and Pedagogical Guardrails

The summit unsurprisingly had a great emphasis on the use of AI in teaching and learning. Attendees shared common frustrations, such as the overreliance from students on large language models (LLMs) like ChatGPT to write essays and solve math or coding problems. I'm going to offer an assumption that we likely all agree that having an LLM return exact solutions to assignments is not a reasonable use of AI for learning. After all, this would be akin to providing exact solutions to students when they come to our office hours. As educators, we guide students through the learning process by identifying alternative conceptions and asking probing questions, among other approaches. So this begs the question, what if an AI could be just as thoughtful of the learning process?

This is where speaker Carter Zenke introduced the catchphrase of the summit: pedagogical guardrails. If we add guardrails to an LLM by means of thoughtful system prompts, for example, we can harness the power of AI to essentially approximate a 1:1 teacher-student ratio. In this method, the LLM will not provide an outright solution to a question, but rather guide the student through the appropriate scaffolding questions to lead them to their own solution. Harvard’s CS50 course, Introduction to Computer Science, has its own approved LLM that students can use to brainstorm, troubleshoot, reflect, and more during their learning experience. You can read more about their solution here.

Other key takeaways from the AI-focused sessions include:

• Duke University has made its guidance on AI for educators publicly available.
• Oregon State University proposes a revised Bloom’s Taxonomy in the age of generative AI.
• Current students are increasingly concerned about the environmental impact of AI.
• AI-detectors are largely distrusted, and instructors are relying on their personal judgment and crucial conversations to enforce AI policies.
• If you are responsible for drafting your own AI policy in your syllabus, check out this mega list of faculty-submitted policies from across the nation.

★ Several attendees were interested in NSF Unidata’s newest eLearning module, Machine Learning Foundations in the Earth Systems Sciences, as an example of how to introduce students to AI/ML without the use of code.

Workforce Readiness and Experiential Learning

In a panel on workforce development and lifelong learning, panelists emphasized the role of higher education for practicing practical skills and creating safe spaces for experimentation. When creating instruction for workforce readiness, consider the skills that are in-demand by employers. What skills are consistently showing up in job ads? In addition to popular technical skills (Python, data science, AWIPS, to name a few), you’re likely to find an emphasis on collaboration, problem solving, and leadership. Consider how you might facilitate the development of these skills in your courses, including how you may employ experiential learning.

★ Consider also suggesting your students apply to the Unidata Student Internship Program, more details coming this winter!

Experiential learning is a pedagogical technique that promotes deeper learning by both exploring subject matter and one’s own learning process. A panel of faculty who have been practicing experiential learning in their classrooms for years shared their perspectives:

• “Simulate professional practice.” Experiential learning does not always mean expensive field trips, and I personally can’t advocate for taking your students out on a potentially dangerous storm chasing excursion in the name of learning. Lucky for us, there are other ways of providing relevant experience for learners. Simulations, case studies, and other projects all fit the bill when they include experimentation and reflection.
• “A commitment to content coverage is holding us back.” Panelists shared that in the modern smartphone age, we shouldn’t be spending time covering content that can easily be looked up. Instead, spend more time engaging in application and reflection of knowledge with students to deepen their understanding.

Concluding thoughts

Nicole and her tablemates on day 3 of the summit

This summit was excellent. The crew at Duke LILE have managed to cultivate a curious, thoughtful, and excited community of education professionals, and I gained a lot just from casual conversations with attendees. In addition to what I am able to describe here in this brief blog post, several other insightful nuggets were shared. You may be interested in knowing about the Last Mile Education Fund, a grant to help students finish their STEM degrees. Or, if you’re interested in tools to help monitor group dynamics in larger-scale group projects, attendees suggested using tools such as CATME. Or if you’re looking for some optimism, check out the resources made available by the closing keynote speaker, Victor Strecher. If you’re interested in attending next year’s summit, I’d love to see you there!