MIT OpenCourseWareWikipedia & MIT Open Learning: A New Strategic Collaboration
MIT OpenCourseWare
Opening Welcome And Introductions 0:06
The session opens with thanks to the MIT Open Learning resource development team and a request for donations to support open educational resources. The host introduces Kurt Newton, director of MIT OpenCourseWare for 22 years, and Andrew Lee, a longtime Wikipedian and author who previously worked with the Smithsonian and now partners with MIT Open Learning. Andrew notes that Wikipedia and OpenCourseWare were both founded in 2001, making this collaboration a meeting of two institutions built in the same era.
Shared Roots And Growing Reach 3:02
Kurt describes OpenCourseWare's growth to 2,600 courses, adding about 50 a year, now consolidating on the MIT Learn platform while also meeting learners on YouTube, where it has nearly 7 million subscribers. He and Andrew reflect on the shared "hacker ethos" of 2001, when the early internet made giving knowledge away freely feel like a radical idea, alongside projects like Internet Archive and Creative Commons.
Finding Gaps In Wikipedia 6:30
Andrew explains that Wikipedia, despite having 7 million articles, still has major gaps compared to the scope of human knowledge. MIT's free course images and materials are being matched to those gaps, identified partly through Wikipedia's own internal warning tags, like "citation needed," which flag articles needing better sourcing or clarity.
Using AI For Matchmaking 9:32
AI tools, including vector search and cosine similarity, help scan Wikipedia's needs against MIT's archive far faster than before, finding conceptually related matches rather than exact keyword hits. The AI does not edit Wikipedia itself; it only surfaces opportunities, leaving final decisions to people, including graduate students brought in to help with this matching work.
Early Examples Of Contributions 13:00
Concrete examples include linking OCW lectures directly from related Wikipedia pages, and adding MIT diagrams, such as space shuttle solid rocket booster illustrations, that Andrew says surpassed even NASA's own materials. MIT's building technology and civil engineering courses also offered comprehensive, methodical breakdowns that filled in disconnected or thin Wikipedia coverage.
Looking Toward Video And Audio 15:30
The conversation turns to OpenCourseWare's roughly 8,000 published videos, averaging about 40 minutes each, often drawn from full 70-minute lectures. Kurt describes efforts to extract shorter, self-contained clips from these longer recordings, citing economist John Gruber's microeconomics lecture, titled "Oligopoly 2," as a recent case where compelling segments could stand on their own within Wikipedia articles.
Planning for Wikipedia from the start 18:01
The speakers describe clipping lecture highlights for YouTube and seeing fifty times the engagement compared to full lectures, showing the value of surfacing the most compelling moments. They discuss whether MIT OpenCourseWare could plan, from the moment a course is filmed, where each piece will live, whether that is Instagram, TikTok, LinkedIn, YouTube, the Internet Archive, or Wikipedia. Both agree this kind of forward planning is already on their minds, since MIT OpenCourseWare and MIT Open Learning are thinking broadly about the many forms knowledge sharing can take, from full courses to quick curious searches.
Wormholes versus algorithmic feeds 21:02
Unlike YouTube, which uses algorithms and viewing history to serve an irresistible next video, Wikipedia keeps people engaged through what one speaker calls the wiki wormhole, clicking from one blue link to another and learning something unexpected fifteen minutes later. This kind of serendipity is described as unique to Wikipedia, and the hosts favor more of this rabbit holing over doom scrolling, noting that AI tools like Gemini are starting to offer similar branching suggestions.
Encouraging contribution to shared knowledge 23:30
With Wikipedia holding over seven million English articles alone, the hosts talk about encouraging both newcomers and experienced editors to keep enriching it. Contributing feels like a positive, activating response to discouraging times, a way to build a civic muscle, and Wikipedia in Paris discussions stressed making sure people know this resource exists and welcomes participation.
Free knowledge under new pressure 26:01
Marking OpenCourseWare's twenty fifth anniversary and MIT's knowledge sharing commitment since 1861, the conversation turns to how media consolidation, paywalls, and polarized information environments make free, shared knowledge more essential than ever. Wikipedia increasingly relies on sources now hidden behind paywalls, and open licensing is called foundational, letting anyone worldwide use this material regardless of ability to pay, while also raising questions about ownership and authorship as AI reshapes how knowledge is adapted and extended.
AI reshaping how people find information 33:02
AI has sharply changed web traffic patterns, with the old search engine optimization model fading as a new generation prefers chatting with content rather than searching for it, something Wikipedia Foundation research confirmed at a recent editors conference. Wikipedia traffic has begun slumping as a result, but OpenCourseWare traffic has not, which the hosts attribute to people wanting a deeper journey rather than a quick transactional answer, making them more willing to follow links into fuller exploration.
AI Traffic and Understanding 35:31
Open Courseware's own analytics show steady traffic plus a sharp rise in referrals coming from AI tools over the past year, even though legal norms around attribution and source linking are still unsettled. One guest notes that AI can speed up research and discovery but cannot speed up understanding, which is why learners still come to OCW's deep content. MIT Open Learning is actively building AI-enabled tutors for its new MIT Learn platform, aiming to personalize the learning journey by tracking whether a student is falling behind, translating classes into other languages, and even swapping industry examples, such as turning a construction case study into a healthcare one. These tutors would also pull in trustworthy outside resources, including Wikipedia, to fill gaps in MIT's curriculum, raising the broader question of training AI tutors well for an informed global citizenry, aided by translation into roughly 200 languages.
Preserving Knowledge Under Pressure 40:01
With institutions like the Smithsonian facing political pressure and exhibits being forced to change, the conversation turns to whether open, volunteer-driven platforms can preserve knowledge that might otherwise be pulled down. The guest points to a long history of citizen participation in science dating to the 1700s and 1800s, and to 2020, when Smithsonian secretary Lonnie Bunch opened the institution's collections for upload to Wikipedia, just weeks before COVID closed the Smithsonian's doors for over a year, making that uploaded content the only accessible version during the shutdown.
Capturing More Teaching on Video 43:30
Recordings made for students during COVID, never meant for public platforms, are raised as an underused record of teaching worth preserving, with video or audio capture of more classes seen as valuable for the future of knowledge. One speaker reflects on wishing great past faculty had been recorded. The discussion highlights that while OCW excels at sharing instructional material widely, creating spaces for learners to actively engage with it has been harder, and Wikipedia fills that gap, with some open education programs already assigning students to improve Wikipedia pages as a way to activate what they've learned and contribute back to the shared commons.
Expanding Collaboration Across MIT Programs 46:30
Introductions mark the final segment, with Peter Kaufman, Andrew Lih, and Kurt Newton continuing the discussion. MIT Open Learning's 192 staff produce far more than traditional Open Courseware, including Universal AI, Universal Climate, short-form Delta team videos on physics and engineering, MITx, and MIT Video Productions. The question becomes how to extend Wikipedia collaboration across all these formats, not just licensed OCW material, with ideas floated like asking faculty to record a short, Wikipedia-ready explanation of a key formula or discovery at the end of each lecture.
Reaching Millions of Learners 50:01
OCW has reached an estimated 500 million people worldwide, and the hope is to enlist that learner community directly, encouraging them to check and improve Wikipedia pages related to what they've studied. One guest describes repeatedly meeting people in Spain, Brazil, and elsewhere who built Wikipedia content based on MIT's open courses, often without MIT even realizing it. Despite being the first major institution to do this kind of work, the team expresses satisfaction that more institutions are now entering the space. The conversation closes by asking whether this MIT-Wikipedia collaboration, still less than a year old, could become a model, with best practices documented so other knowledge institutions can learn from and replicate it.
OE Global Conference at MIT 53:30
MIT is hosting OE Global 2026, the largest worldwide gathering of people working in open education, with hundreds attending in person and online, reflecting a shared mission of making knowledge available to everyone.
Escaping Digital Feudalism 54:58
The speakers describe modern subscription culture, paying rent to many streaming services, as digital feudalism, and argue that open resources like Wikipedia, Internet Archive, OpenStreetMap, and OCW offer free escape from that trap, forming a commons worth protecting.
Treating Knowledge as Public Good 55:32
Holding knowledge as a public good means continually investing in it. An Open 2030 working group has explored adapting this work for community colleges and historically Black colleges and universities, many of which are embattled, hoping for broader systematic collaboration.
Closing Thanks and Acknowledgments 57:30
They call for a federation of groups working for knowledge in the public interest, then thank Phoebe Ayers, Steve Carson, SJ Klein, Kurt, Andrew, funders including Hewlett, Sloan, Arcadia, and Templeton foundations, and MIT alumni, before closing the episode.
AI-generated summary. It can be wrong or incomplete - check anything that matters against the original.
