Skip to main content
SearchLoginLogin or Signup

Meeting #8, April 29-30, 2024

@ MIT Open Learning

Published onApr 09, 2024
Meeting #8, April 29-30, 2024
·

Meeting #8, April 29-30, 2024

Agenda

Monday, April 29

3:00 PM

Meet at MIT Press Bookstore

https://whereis.mit.edu/?go=E28

3:30-4:15 PM

Begin MIT-IBM Watson AI Lab tour with Sarah Schwettman

4:20-4:50 PM

Lightboard studio tour with Jim Cain

5:00-5:15 PM

Building 55 (Climate & Sustainability space) tour

6:30 PM

Drinks and Dinner

Tuesday, April 30 – Open 2030 Working Group meeting 

9:00 AM - 3:30 PM 

MIT Tang Center - Building 51 – Room E51-385 

https://whereis.mit.edu/?mapterms=E51-385

9:00 AM Breakfast & coffee

10:00 AM Welcome 

The Spring meeting of the Open 2030 Working Group was held at a time where AI, climate, geopolitical conflict, and attacks on higher education are widely shared concerns and important settings for deeper collaboration.. It was held as a hybrid session on April 29-30th, 2024, at MIT Open Learning in Cambridge, MA. The meeting kicked off at the MIT Press Bookstore on the 29th with tours around MIT campus, including an introduction to AI and climate initiatives of the Sara Beery research group at the new MIT Schwarzman College of Computing, a look into the MIT Lightboard recording studio, and a tour of the MIT Climate and Sustainability space.

10:15 AM Generative AI and open

Sarah Schwettmann, MIT CSAIL & MIT-IBM Watson AI Lab  

MIT CSAIL’s Sarah Schwettmann presented a comprehensive exploration of new and upcoming AI tools and how they can amplify open education. What does it mean for AI-based educational resources to be open? The answer can be summarized in 3 aspects: transparency, accessibility, and customizability of AI. Transparency relates to how much of the inner workings of the educational AI tool is available knowledge, as most of these tools are built upon opaque proprietary models which allows for strictly internal research, high risk control, low auditability, and limited perspectives on the model. Examples of models that are fully closed include Imagen by Google, whereas GPT-4 by OpenAI can be used with hosted and cloud-based access. On the other end of the spectrum, BLOOM by BigScience is completely open, which means the community can help retrain the model on better data sets through open auditing, and explore biases, paid placements, and manipulation. Accessibility is the second crucial aspect to effective educational AI, especially since open models are very expensive to run. The cheapest models to use are currently proprietary (like GPT-3.5), commercial or off the shelf, but they’re heavily subsidized to build familiarity and lock in. The bottom line is that open options should be the easiest and cheapest for students to use. Lastly, the AI models that will succeed are those that are customizable to meet the specific needs and preferences of individual learners and diverse audiences. Tools that can recognize patterns in user interactions and performance data will create the most effective educational experience.

11:15 AM Break 

11:30 AM Climate and open

MIT OpenCourseWare director Curt Newton presented a report of MIT’s recent climate change work oriented around “The Climate Project at MIT'.”. The main goal of The Climate Project, as Newton explained, is to “do bigger things faster” using MIT’s talent and resources to change the trajectory of our climate. The Institute is currently recruiting faculty leaders to fulfill their six missions: 1) Decarbonizing Energy and Industry, 2) Restoring the Atmosphere and Protecting the Land and Oceans, 3) Empowering Frontline Communities, 4) Building and Adapting Healthy, Resilient Cities, 5) Inventing New Policy Approaches, and 6) Wild Cards. The Climate Project at MIT will also support climate education activities that empower students to take interest in climate justice, such as the Responsible AI for Social Empowerment and Education (RAISE) initiative for ages 8-18, MIT Climate Action Through Education (CATE)for high schools, and the Climate Justice Instructional Toolkit for undergraduate students. 

This topic was followed by a presentation from Daniel Erasmus, creator of ClimateGPT, a new Large Language Model (LLM) synthesizing interdisciplinary climate change information for researchers, policymakers, and business leaders. The tool can answer a range of questions while also providing data that shows that climate change affects people just as much as it does the environment. Drawing on climate data to analyze and predict climate events, ClimateGPT offers innovative solutions, engaging educational content, and resilience planning to make climate change knowledge more accessible. And unlike its counterpart ChatGPT that uses copious amounts of water and energy to run, ClimateGPT has aligned with its values by using fully renewable energy to power itself. Visit https://climategpt.ai/ to try it out.

12:00 PM Funding the Commons

MIT Open Learning’s Peter Kaufman addressed the importance of funding the commons in preventing mis- and disinformation and preserving the digital Commons as a public good. The World Economic Forum calls mis- and disinformation the biggest global risks for the next two years, even more so than war or extreme weather. This calls for an urgent need to fund the commons and keep information, as well as educational resources, as open as possible. Funding for platforms dedicated to open access like Wikimedia and Internet Archive must continue. In April, 2024, the Wikimedia Foundation called for UN Member States to include three commitments in the Global Digital Compact to allow community-driven projects to thrive: 1) Protect and empower communities to govern online public interest projects, 2) Promote and protect digital public goods by supporting a robust digital commons from which everyone, everywhere can benefit, and 3) Build and deploy Artificial Intelligence and machine learning to support and empower, not replace, people who create content and make decisions in the public interest. They are specifically calling to edit the Zero Draft to finally recognize the public value that Wikipedia has as the only community-governed, nonprofit model in the top-ten most-visited global websites. The fate of the organization also will be discussed this September at the UN Summit of the Future. Find MIT guides to mis- and disinformation here and data here.

12:30 PM Lunch

Updates from the field 

1:30 PM Textbooks and syllabi 

On the topic of public goods, the following presentation reported on the progression of the Open Syllabus Project by director Joe Karaganis. He recently published an OER Adoption Update in April 2024 upon the release of new data showing significant growth of open access/open educational resource textbooks and monographs since 2013. Compared to the 1 in 400 classes using OER textbooks in North America 10 years ago, it now averages 1 in 80 classes, and 1 in 40 classes in 2-year schools. Due to the decentralized nature of US higher education, OER adoption varies based on state, system, and individual school policies. California, with its well-supported state-level program, leads significantly in OER usage: over 1 in 40 classes use OER titles, increasing to 1 in 25 at 2-year schools. On the other hand, states like Georgia, Florida, and Illinois remain far below the average for OER adoption. Math and computer science have the highest percentage, with 1 in 25 math classes using an OER textbook. On a global level, adoption is tilted toward monographs, with the UK showing rapid monograph growth and very little textbook usage. Overall, the growth rate for OER titles will continue to be sustained worldwide and the Open Syllabus Project remains successful in lowering the cost of higher education.

2:00 PM Discussion on open education’s responses to the assault on higher education

The Spring 2024 Open 2030 Working Group session concluded with a discussion led by MIT Open Learning’s Senior Associate Dean Christopher Capozzola to answer the question, “In 2024, how can leaders and stakeholders in the open education movement contribute to positive outcomes during a time of political division and systematic attacks on higher education in the United States and around the world?” The vigorous discussion was held off the record. 

2:45 PM Closing discussion & Fall 2024 event planning

Comments
0
comment
No comments here
Why not start the discussion?