AI governance, literacy, and the power of connection: Three themes from J-WEL Week 2024
As the impact of generative AI technology on education continues to dominate headlines, the MIT Jameel World Education Lab (J-WEL) and its member institutions are exploring this topic head-on.
They recently gathered for J-WEL Week, the lab’s flagship in-person event that is themed each year around a topic that is top-of-mind for member institutions. This year’s topic: generative AI in education. After a week of shared conversations, collaborative workshops, insightful presentations, and in-person demonstrations from the MIT community designed to deepen understanding and fuel collaboration, three key themes emerged.
University governance of AI is taking shape in real time
Several sessions during J-WEL Week addressed the theme of governance, as all member institutions grapple with what policies to implement (or not) regarding the use of generative AI. Not only is there concern that students will use the technology to bypass real learning, educators are uncertain about the level of understanding they need to have of generative AI tools.
During day 1 member updates, two universities shared how they are tackling the question of AI governance. Professor Melchor Sanchez Mendiola spoke about the GenAI in Education working group that was created at the National Autonomous University of Mexico (UNAM) to explore this topic. The working group, including engineers, educators, sociologists, computer specialists, and distance education specialists, is proposing guidelines for UNAM that take into account different perspectives across the university ecosystem.
Similarly, Professor Fabio Cozman spoke about a workshop hosted in 2023 by the Center for Artificial Intelligence and Machine Learning at the University of São Paulo (USP). The group met to discuss the impact of ChatGPT and other generative AI models on education, and ultimately produced a set of recommendations on how to study and discuss AI within USP.
MIT Schwarzman College of Computing Dean Daniel Huttenlocher focused on AI governance from a macro lens during his presentation on day two. Huttenlocher shared a brief history of AI, contextualizing the rapid advancement in the large language models (LLMs) of today. He also cautioned against creating AI policies in a vacuum, instead recommending that AI be worked into existing policies.
All week, presenters and attendees alike agreed that a perspective shift of generative AI is needed. Rather than looking down on AI as a way to circumvent original thought and work, they are thinking of it as a tool to spur creativity and enable iteration and feedback for learners in all disciplines.
Baseline AI literacy is necessary for educators
The question of AI literacy was also prevalent: how much expertise does an educator need on generative AI? Members agreed that a baseline understanding of AI tools was necessary for both educators and students, drawing metaphors to calculators and word processors. These were similarly disruptive educational tools that all educators ultimately needed working knowledge of, and attendees agreed that generative AI will be no different.
Professor Eric Klopfer summarized this viewpoint in his presentation outlining the mission of MIT’s Responsible AI for Social Empowerment and Education initiative (MIT RAISE). Klopfer defined AI literacy as “a set of competencies that allow students to be informed users and responsible producers of AI,” stressing that understanding the fundamentals of AI is important for making informed decisions and participating in our civic, work, and social lives.
Generative AI in Global Education Lead at J-WEL, David Dixon, tested attendees’ AI literacy with a hands-on workshop focused on writing better prompts and building custom GPTs. This allowed members to see real-life examples of how working knowledge of ChatGPT, Copilot, and other LLMs can be a tool in an educators toolbelt, and test out creating their own custom GPTs.
In-person gatherings foster unique connection to the MIT community, and each other
J-WEL’s member institutions regularly report an intangible value in gathering in-person for J-WEL Week. The open dialogue, workshops, and impromptu conversations during coffee breaks allow members to find common challenges and opportunities amongst each other, while the talks from faculty, students, and staff uniquely connect attendees to MIT’s culture of excellence in education. The agenda was designed with these insights in mind.
The final two days of J-WEL Week, members participated in a design thinking workshop created by J-WEL Faculty Director Anjali Sastry and J-WEL Research Scientist Joe Doiron. Attendees co-created toolkits and recommendations on AI governance, literacy, prompting, and using AI tools with students, with the goal of taking home a set of shared resources to introduce to their home institutions and apply outside of J-WEL Week.
Paula Elksne, director of the Education Innovation Lab at Riga Business School, summed up in Thursday’s closing remarks that J-WEL Week is a “good mix of hands-on work and hearing from others [about AI].” She added, “I’m now thinking more of how we can keep each other accountable to what we’ve learned and support each other as we go to implement [the learnings].”
...
For additional resources, explore MIT’s collection of Exploration of Generative AI papers and the Sloan School of Management’s Generative AI for Teaching and Learning resources.