Wikipedia & MIT Open Learning: A New Strategic Collaboration: transcript

Thank you for joining us for this Open Learning conversation about Wikipedia and MIT Open Learning, a new strategic collaboration. Thanks also to our small, but mighty Resource Development team here at MIT Open Learning for finding all the ways to organize this online event in a very busy week at MIT.

And this RD team is so good, they make me recite the following. Before we dive in completely, I'd like to acknowledge that so much of what we do is possible because of support from our friends at charitable foundations, companies, governments, other knowledge institutions, and I can imagine from many of you who've tuned in. If you're able to, please consider making a donation to help ensure that high-quality educational resources remain open to all the millions of people around the globe who look to us and initiatives like this one for trustworthy, publicly available knowledge and information.

Joining me today-- Curt Newton, director of MIT OpenCourseWare. He has been with OpenCourseWare for 22 years, since February of 2004, approximately 8,283 days. And Andrew Lih, a lifelong Wikipedian who has worked with knowledge institutions around the world, who wrote the great book on Wikipedia's early years, and after helping the Smithsonian Institution advance some of its objectives similar to ours, is now working with us at Open Learning.

Andrew, I will start with you. Why are you such a pleasure to work with? And why have you been so good to us? [LAUGHS] Well, I think until I worked with you, I didn't realize that Wikipedia and OpenCourseWare are the exact same age. So both were started in 2001. And I think it's just perfect that we're working together in this way of bringing knowledge to the world in ways that are uniquely suited to this new-- the new way that we are learning and accessing content online.

So it's been great working with you folks. Thanks. Curt, you've been with OCW for these 8,000-plus days. And you've seen it grow to now 2,600 courses, with 50 new courses each year. We're consolidating. We're syndicating our knowledge offerings around the place, consolidating some of them on this new universal platform, MIT Learn.

But I know that you've been keen on meeting learners where they are. Our work on YouTube, which you've led-- we can count ourselves as the largest educational contributor there, definitely one of them. And we have almost 7 million subscribers. Wikipedia is also where the people are and increasingly, where the machines are.

What do you have to say about the potential for working with Wikipedia in a similar way? I think the potential-- the opportunity is staggeringly large and expansive. Referencing Andrew, you're talking about sharing this birthday, this 25-year-- of 25 years ago. There was something in the atmosphere, the environment in those early 2000s years, that I think helped give rise to OCW and Wikipedia, Internet Archive, and a few other things around that time.

Creative Commons. And Creative Commons, yeah. And we've been really busy these last 25 years in our respective lanes, if you will. Some of those lanes are bigger than others. I think we're actually entering into a similar moment of there is something really big and expansive in the air that calls on us to work together in new ways.

And I feel like there's something in this collaboration that has us returning to some really important mission-driven roots to reassess where things are at and take it from there. And I think we'll probably get into that in the course of this hour. What are some of those things? What are some of those things? We should actually open up the discussion to provide some details about some of the things, Andrew, that you've been spearheading with us and our whole OCW team has been so excited about.

Could you say some of-- just a word about some of those achievements and where things are today? Yeah. I think as Curt said, I think there's-- the shared lineage, I think, between Wikipedia and MIT OCW is-- when I wrote my book about Wikipedia, it's about this hacker ethos. I think that same spirit was in 2001, where-- OCW just celebrated their birthday recently. And there's a great video that you folks produced. And it was great seeing the interviews with the professors and folks who started this whole thing.

And it reminded me a lot of Wikipedia's beginnings of saying, the internet has allowed for everyone to access content on websites, and why don't we have this radical idea of giving things away? And there's power in giving things away. And with such deep resources that you have at MIT and with Wikipedia being the number one reference site in the world, I think that was a great, as you said, collaboration to have. So a lot of the diagrams and the course materials and just world-class content that MIT has-- even though Wikipedia has been around for 20-some years, there's still lots of gaps in things that could be contributed to.

And Wikipedia's need for those things is nonstop. And MIT just had some amazing resources to bring to the table. And that's why even when I was with the Smithsonian and with the Met Museum and working with Wikipedia, there was always more stuff to add. Wikipedia's 7 million-some articles. But that's still quite small compared to the expanse of human knowledge. And I think we've done a pretty successful project right now with uploading some of those free images from MIT courseware and identifying places where Wikipedia needs that help and finding those gaps in knowledge.

Could you both speak to the process that we've undertaken to identify where those gaps are and maybe highlight some examples? I know that during Wikimania 2026 in Paris, there was an extraordinary presentation with all kinds of slides. Hopefully, we can upload those slides to our audience today.

Can you cite some examples of the work that we've actually been able to develop together, and more specifically, the process of identifying where and how MIT OpenCourseWare and Wikipedia can complement each other? Yeah. For folks who don't know, Wikimania is our annual conference we have in the Wikimedia community. And it's pretty big this year in Paris. So Peter and Curt were able to see how the insiders work on Wikipedia.

And it's a pretty amazing, bizarre and-- mixture of folks working on all kinds of different things. So I think the primary thing that we want to do when trying to work with-- when Wikipedia tries to work with outside or other collaborations is that we want to make sure that there's a respectful and harmonious type of collaboration. So in this instance, we tried to find areas where Wikipedia has a demonstrated need or desire for better content. So some of you using Wikipedia might know that we have these things called warning tags or labels inside.

So you might see a "citation needed" or "this article needs more sources" or "this could be rewritten to be clearer." And what's one of the great things about Wikipedia is it's very transparent about its flaws. As much as we think it's a great resource, it's the number one reference site, it's quite open in saying, we could use some help here. And it's those situations where we can very quickly and programmatically go through and find out what articles are high-impact, but have a high need for improvement and then, knowing the corpus of what MIT has with OCW, find ways to connect those two, which we call matchmaking, to make sure the opportunities for contribution match what MIT has.

I'd amplify a few things that have really stood out for me in this process, that the Wikipedia community is part of the culture and the process there-- has those flags and that self-awareness, if you will, about the state-- the current state of the content-- enables anyone with a large collection of material, like OpenCourseWare, to say, how might this map to what we can offer there? And we could keep humans busy for many, many, many years going through and assessing all of that stuff.

I think another key part of what we've opened up through your work, Andrew, in particular is bringing some AI tooling to allow us to do this at scale. And the AI tools are not making the final decisions in any way. But they're taking us from these two enormous piles of information, what Wikipedia needs and what OCW has got, and doing some really helpful initial matchmaking.

At the end of the day, it's going to take human insight to actually make some of those decisions. But it's been really gratifying to see the amount of overlap and opportunity that sits there. And so that's really what we're now starting to lean into to try to activate that. I keep walking into these meetings, and someone says something about cosine similarities. And I turn around and walk out of there. Me, too. I have no idea. But apparently, we're working-- our vice provost for Open Learning has brought graduate students into this mix who are experts in cosine similarities, who are experts in-- how are you working together with graduate students from MIT?

I think one of the most exciting things right now is that in this age of AI, our algorithms and our ability to find things and identify them are so much better and faster than before, as you said. And until you see it in action, you may not realize it. So again, we're not using AI to edit Wikipedia at all. We're trying to use it to find the opportunities better. So in the area of what we call-- what you call cosine similarity-- Not me, him. --or embeddings, or vector search-- you might hear some of these terms in the AI field-- it's trying to go one better than what you're used to seeing with a Google search.

So you're used to saying, I'm searching for "horse." But if you are looking for different things, like ponies and mustangs, a search for "horse" is not going to find those. But in this day and age, with vector search, with cosine similarity, those types of algorithms, you will find all different types of horses and ponies and mules and things like that because it knows the meaning of that word. And that's the secret sauce that we have today that we didn't have 25 years ago, that our ability to do those types of matching are so much better than before.

And that gives us a lot more high-quality types of opportunities. So MIT "OpenHorseWare"? [LAUGHTER] No. Sorry, sorry. Are we live? Yeah. We did practice that, by the way. An example of MIT OCW content that's now in Wikipedia is what? You mentioned-- I'll cite a couple of things. One of the first places we leaned into is just Wikipedia pages where there's a really pretty direct, obvious match with either a lecture in OpenCourseWare, or even an entire course.

And so putting an external link-- a reference, if you will-- from that Wikipedia page to the OCW material-- if you're interested to learn more, go here-- and offering that up also to the editing community-- if you want to dig in and do more refinement, this is a great place to look. That's one of the most obvious and direct low lifts. And Andrew mentioned also bringing some images and things in and really starting to open that up as well.

So you did this great one on the space shuttle booster that you found. What was that? So it's interesting that some of the things that MIT is so good at in terms of STEM classes-- so if we have things related to architecture, building technology, civil engineering, aeronautical engineering, all these things have classes with great resources. Some of these are even, I want to say, better than NASA. I used to work at NASA for one year. So they are unique in that some of these diagrams were really better than anything that was in Wikipedia.

So for the solid rocket boosters of the space shuttle or the space shuttle construction, some of the diagrams that were in the OCW classes were better than what was put up there in terms coming from NASA. So those were put in some very high-profile articles. Building technologies-- just having classes at MIT that laid out all the different types of building construction were really valuable because it was a methodical breakdown of a topic. Whereas Wikipedia might be good in one article here, one article there, it's not always comprehensive across a set of disciplines.

So those have been very valuable. And I'll call out-- we're really, at this moment, trying to lean in as much as we can into media enrichment of what are otherwise often fairly text-heavy Wikipedia pages. It feels like there's something in particular-- a value that we can try to add from our collection into Wikipedia to make-- just to make the content work more effectively for all the learners it's trying to reach. There's an enormous database-- this came up at Wikimania-- of images and videos and sound that's part of OpenCourseWare.

And I've seen you both analyze this stuff and imagine how that material can match to spaces that are now open in these Wikipedia articles. Is there a future for doing more video, for example, more audio? I remember in Paris, it came up that if Wikipedia were to start today, it would probably be a very video-rich environment.

And here, I don't how many-- I how many days you've been here. But I don't how many videos there are in OpenCourseWare, how many minutes of video. Actually, we have that number. Quick sort of-- have an estimate. There's about 8,000 videos that we've published. And they're, on average, about 40 minutes. So I don't know. Do some math there a little bit. I can't do that. I work here. In OCW, we've been experimenting with-- a lot of these videos are from a full 70-minute classroom lecture.

And those are fantastic in their own right if you're ready to sit down and watch a 70-minute video or engage with it like that. But also, within those longer form videos are these incredible crystalline nuggets that-- to the extent that we can identify those and allow them to travel independently, so much the better. A very recent example-- we've been working a lot with an economist here at MIT, Professor Jon Gruber, who teaches Introduction to Microeconomics.

In his lecture-- I think the lecture is titled "Oligopoly II," that attention-grabbing title-- are several fantastic embodiments of why this stuff is interesting and useful and is important. And we've put that out on our YouTube channel as a test case and gotten, like, 50X the engagement on those clips. And the full-form lecture got plenty of engagement, too. So it's indicating the opportunity to-- again, identifying the most salient, more easily engaged with content and putting it to where the people are.

And I can't think of a more important place where people are than Wikipedia. And so we look forward to trying to channel some of this stuff in. No doubt. And it would be, I think, of primordial interest to the audience on YouTube tonight to see some account of how this progress is coming. So either an article, a video, more of these conversations, I think, with actual examples-- is it conceivable that in the future, when OpenCourseWare, these 50 courses that get produced each year-- part of the production process could actually look not only at where OpenCourseWare will go on OCW's site and on MIT Learn and were pieces of video might go if they go on Instagram and TikTok and LinkedIn

and what have you, but also YouTube? And is it conceivable to imagine now thinking about the Wikipedia audience when the course starts and actual video shoots are commencing? Is that a-- More than conceivable. It is on our minds. Is it on our minds? Yeah? Yeah, yeah. So has it been the case that we've done this with-- I think with YouTube some-- automatically, in fact-- and the Internet Archive, which is another extraordinary ally in this space today.

But is it-- it's on our minds? Yeah, yeah. At OpenCourseWare and across MIT Open Learning, there are various projects that are thinking about this wide spectrum of the forms in which we want to share knowledge out with the world. And some people are up for what feels a full course experience. Some people have something very specific in mind that they're driving for. Some people are just curious and poking around. And we want to be able to meet all of those needs.

How does it work on Wikipedia? We were talking a little bit about this earlier today at the hackathon-- Edit-a-thon. --edit-a-thon that you led with Phoebe Ayers, the great Phoebe Ayers from MIT Libraries. We'll get another shout-out later. On YouTube, you watch a video, and you're immediately served up something else.

It's over. And you read a Wikipedia article-- obviously, it doesn't track your-- but how do we make sure that people continue to dig in to Wikipedia with the same kind of self-directed force as the algorithm is propelling people to go into new minutes of sound or video on social media? I think the model of engaging with Wikipedia and keeping people engaged is quite different than the YouTube version, where they try to use an algorithm and all the history they know about you to surface a video at the end that is too juicy not to click on.

But Wikipedia-- I'm sure many of you have tried it. And you've gone down the Wiki wormhole. So you read an article. And you click on a blue link into something you may not know about. And you click and click, and you accidentally learned something new 15 minutes later, which I think is wonderful. That serendipity and that wormhole is something that I think is unique about Wikipedia. What that looks like going in the future, it will be interesting to see. But I think it's something that not many other sites replicate. I'm sensing something starting to take shape out in the world where that experience-- you call it wormhole.

I call it rabbit holing, where other people-- you have-- sometimes, you have to go outside the bounds of a particular application to really get into that. But people are starting to try to put a more intentional experience on what it's like to just follow your curiosity from thing to thing to thing to thing and be able to keep track of where you've been and get back to it. And I think there's something in the world we're heading into where that's, I think, increasingly important.

One of my new favorite phrases is "more rabbit holing, less doomscrolling." That's good. And you see that in some AI systems now. I think to my perception, Gemini does it more than anything else. They tell you something at the end like, do you want me to tell you about X, Y, or Z? So it's kind of doing that same thing in the AI domain. Yeah, yeah. How do we encourage people who are on Wikipedia, either experienced Wikipedians or novices, to engage more with the-- we're putting material in the Commons.

We're putting material elsewhere. How do we encourage people to engage with that? We've talked about gamifying some of this stuff, or you have. And are there particular ways of just pointing to it and encouraging either the MIT community, the alumni who are watching us tonight, or others who care about trustworthy knowledge to continue to work with this to make sure that it enriches all of these 7 million-plus articles?

Is that English-only? 7 million is English. Most of the other languages are somewhat smaller than that. But still, 7 million is pretty big. Yeah, that counts. That's almost as many days as you've been at-- [LAUGHTER] Yeah, yeah. I think there's something really powerful that we could be working together to try to tap into in the grand scheme of things out in the world. A lot of people are distressed about the way some things are going out there.

And you could feel discouraged. But for people who are looking for ways to get activated and do something that feels positive, I can hardly think of something that's more positive than contributing some knowledge back to the Commons about something that you care about. That is so true. If you don't want to go knock on doors for electoral things because that's-- do this.

Build that muscle. I think it's especially important right now-- and I heard a lot talk about this at Wikimania in Paris, about just making sure that the world knows that this resource is there and ready for you to participate in, to contribute. That's beautiful. And that's really important. Build that muscle. Yeah, yeah. That's great. Let's talk about that a little bit. You've mentioned that this is the 25th anniversary year of OpenCourseWare.

We had a celebration. Andrew, you came up for it, in fact, in April. President of MIT, Sally Kornbluth, spoke so eloquently at it about OpenCourseWare and MIT's commitment since 1861, even longer than you've been here. I'm going to keep-- Keep riding that-- Keep going. Keep riding-- since 1861 about sharing knowledge. And it seemed like 25 years ago, OpenCourseWare started in part as a reaction to the commercialization of the internet and education on the internet.

It was decided-- president of MIT, together with a couple of foundation presidents-- to make this knowledge freely available and not available for sale. We've talked about the Columbia alternative prior to going on the air. And at the time, Wikipedia was starting, too. And it seemed like this kind of work was a good thing to do. Now it seems with the things that are going on, as you've put it, Curt, it's essential, if anything, the consolidation of some knowledge providers from the commercial sector in science, medicine, essential vectors of-- the publishing industry's what they call media capture-- today,

CBS is not owned by the people who used to own it. Twitter is something different. And it's not owned by the people who start-- a lot of the information media that we were counting on or thought we might count on all those years ago-- very different. And the country is perhaps even more polarized. That may be debatable.

Various countries are even more polarized around information that they're each consuming, but not necessarily consuming together. This may be a leading question. But don't you think that this-- creating kind of a common knowledge base of the kind that OpenCourseWare has been building from MIT, from MIT, but especially the kind of common knowledge base that Wikipedia exemplifies, is more important now than ever?

You can just say yes. No, but what would you say about that? I'd say everything that was true in 2001, when both these projects started with both Wikipedia and OCW, is even more true today. And in fact, we're seeing more and more of the sources that Wikipedia uses starting to be put behind firewalls-- or pay walls, I should say, not firewalls, but paywalls, so that CNN, most of these sites that were free for most of Wikipedia's history are now subscription-only or only access to certain parts.

And then now with a lot of the activity with AI, a lot of sites have the "you must prove you're human" before you get to take a look at this stuff. So there's potentially more barriers than ever to get to information today, which makes projects like these even more important. It's been said that the truth is paywalled, but the lies are free. This is a great line from the journal Current Affairs. Wikipedia is free.

Sorry. Isn't OpenCourseWare free, too? It is. I knew there was something going on here. And there's got to be some-- you were talking about alliances to strengthen here, especially given what's going on in the world. And I think the free aspect of that is foundational and really, really important so that literally everybody around the world can tap into this important stuff, regardless of ability to pay.

But also, I think really important is open licensing and how that is evolving in the world we're heading into around AI. It's one thing for a professor at MIT to put out a perspective on a topic that they're deeply immersed in and that makes sense for the context in which they're working. But it's also really important that people recognize that they can adapt, they can be inspired by and extend that work.

And I think in this moment, another thing that's really important is perhaps taking a different look at what this ownership of ideas and senses of exclusive authorship mean. And the pendulum is always swinging on that a little bit. Creative Commons licenses gave people the ability to relax their grip on some of that. And what AI models are doing is just taking another run at that space into some places that were maybe left ambiguous before.

The way in which we take in knowledge, are inspired by, extend, add our own thoughts to it I think is a really important space to be working in, and especially around this case of what's true, what's real? Well, that's going to depend on some of the context. This locality here is going to need to adapt what I heard to make it relevant. So I think free and open-licensed together with this new thing that's opening up around how AI is working is a big space.

Sorry, I went very big on my answer there. That's awesome. Very big is what this is all about. This just in, this AI thing-- it's going to be really big, I think, really big. Have you heard about it, the AI thing? So the AI thing-- We got 30 minutes in before we really got into it. No, this is an achievement, actually. And you had to mention it minute 29 or whatever. No. But the AI business-- how is AI affecting Wikipedia?

And how is AI affecting OpenCourseWare? Well, I think AI is affecting the web in general. We were talking about this before. If you look at just traffic to websites in general, they've declined quite a bit in the last two years where with the debut of ChatGPT, it has changed the paradigm. Some people will say the whole link economy is gone, that we used to depend on Google to drive content to things or getting your SEO up on Google was the only thing that mattered if you're a business or a site.

That has all changed now. We were just recently in a conference of Wikipedia editors. And we got some-- we watched some market research videos that the Wikimedia Foundation was putting out there and saying that the overwhelming desire of this new generation is to interact by chat with content. So it's not a search box. It's not-- By chat. --by chat. So that was one of the things that people had surfaced, saying that they want a chat interface to Wikipedia.

So the same way they would go with ChatGPT, they wanted to interact with Wikipedia in some chat interface. And that's what this generation is accustomed to, the same way that we thought Google Search was revolutionary. To these young next generation, the chat is revolutionary. I don't think Google Search will go away. But the paradigm is massively shifted so that we are seeing last year or two, traffic to a lot of websites, including Wikipedia, start to slump. So that is certainly something to be concerned about, and having to react to that.

It's interesting. For OpenCourseWare, we haven't seen our traffic slumped. Good. And I'm very curious to keep digging into what some of the reasons might be for that. What is it that has the traffic continuing to come to OpenCourseWare when it might be tapering off to others? I have some hunches along the lines of in my experience using AI-summarized search-- is it-- if I'm looking for something that's fairly small-scale, transactional in nature, I'm probably satisfied with what I get from Google, the AI summary.

But if it's something that's more substantial, this is a journey I want to put myself on, I'm happy to take that summary. But I'm much more inclined to follow through and explore those links. That's how my mind works. I don't know about others. But our analytics for our website is bearing this out, that our traffic has stayed steady. And in fact, we're actually seeing over the past year this enormous growth in referrals to the OpenCourseWare website from AI tools.

And so while some of the legal requirements about how attribution and source linking are still in flux, it's happening. And we're glad for that. I think you might be benefiting from a great observation I heard someone say about AI saying that AI can rapidly accelerate research and discovery, but it cannot accelerate understanding. I think that's what they come to OCW's deep stack of content for-- is learning and understanding.

And no super-speed AI can accelerate that for them, per se. Any of you who fill out the OCW surveys that we put out periodically, this will definitely be part of the next survey. We want to hear from you what your experience is. Well, I know at Open Learning, we like meetings. We like our hour-long meetings in particular. And I've sat through enough of them to know that on the docket now is a concerted effort to think about how to customize and personalize both courseware and the learning journey in such a way that if you go on to our new site, MIT Learn, and you are taking a class, you might have an AI-enabled tutor to go along with you to that class, suss out whether you are in any way falling behind.

You may also be able, if you're in another language, to have that class generated in your language. You may actually have a-- the ability to request that a class that uses certain examples from certain industries have an AI-enabled translation effort such that it is now a class that refers to cases in health care as opposed to cases in construction.

Is this something that OpenCourseWare production is thinking a lot about? Are we actually thinking about also not just human learners now, but machine-- teaching the machines? Is that a thing? I know we're at MIT. I got to ask. There's a few questions in there. I will say we are absolutely not just thinking about, but actively testing and building and prototyping these tools that will give you a more personalized journey.

And my hope is that they're not only trained really well on our materials, but are also able to pull in additional things. As complete as our representation of the MIT curriculum is, there are still some gaps and things like that. And to be able to draw in other good, trustworthy, open resources I think will be an important part of creating these good learning journeys. So-- including things like Wikipedia showing up in there. I'm not even sure where to go with your last thing about the teaching machines how to learn more effectively.

I'm wondering about that. Do actual faculty need to think about not just the students in the classroom, but theoretically, if you put the course online, what the machine's going to make of it? To the extent that people are looking for AI to be their tutors, where are tutors trained from, I think in the grand scheme, if we're committed to having an informed, engaged citizenry around the world, it's in our interest to have these tutors do a good job.

Well, I think that there's the multilingual aspect as well. When the AI ingests the courses here in English, they can now be made available to folks in 200-some languages. So that's something that's not nothing. It's also not nothing on Wikipedia, by the way, when I've seen the beginnings of our work get into articles and we-- they get translated quickly into sometimes 22 languages. It's an amazing community.

People on Wikipedia appear never to stop. [LAUGHS] Let me ask a question because you were at the Smithsonian. I know we've talked about horses and stuff. But the elephant in the-- all of these knowledge institutions are under tremendous amount of pressure now. I think it's fair to say that. And the Smithsonian has borne the brunt a lot of pressure on knowledge in Washington with exhibits forced to change and even the current director of the Smithsonian now.

Stepping down, is there a way to look at the Commons and think, here's a place where knowledge can be preserved and bolstered by all of these volunteer editors who are busy contributing and editing and inserting and deleting and then re-ed-- Smithsonian material such that it can never be pulled down from the web in the same way as it gets pulled down from an exhibit, or if there's some injunction to stop teaching a certain subject at a certain knowledge institution, somehow, the teachings on that subject will be preserved in the Commons in a way that will render them more persistent?

That's a good question. I think it's important, especially when I was working with the Smithsonian on these open-access projects, that there's a deep history of volunteerism and citizen science going back to 1700s, 1800s. So a lot of the early work on fossils and things like that were all done by ordinary folks sending in things to the Smithsonian, and to other places as well. So we've always had this participatory dynamic when it comes to knowledge for the public interest.

So with the Smithsonian, when we've been working with them for a long time on trying to get their collections more open-- and finally, in 2020, Lonnie Bunch, the current secretary, was instrumental in getting this done at the Smithsonian to make sure that any and everything available on their website could be downloaded and uploaded to Wikipedia. And that had a strange intersection with the fact that our ceremony at the Smithsonian, which he hosted in February of 2020, was literally weeks before COVID hit full-blast in the United States.

So if you folks remember, shortly after that, the Smithsonian closed their doors and did not open for more than a year because of COVID. But thank goodness we were uploading content from the Smithsonian to Wikipedia. And you could still experience the best of what the Smithsonian had to offer because we had uploaded it and it was made available in Wikipedia. And that was a great example of how even with the Smithsonian's doors closed, having that content on an open site like Wikipedia was useful and helpful. And I think that's a great lesson for everything that we're doing-- is preservation, access, and discoverability.

Those are all things we want to foster. The record, too-- I know we had a visitor last week, a faculty at Dartmouth, who came and gave a small lecture-- big ideas, small lecture-- about an online platform he's building. And one of his friends, a faculty member here, was asking, what about the recordings that were made of my class during-- do you remember this-- during COVID? These weren't necessarily recordings that were made in order to put them on the-- on YouTube, where folks are watching now.

But they were made to share with students in a semi-private or local area and-- I don't know what we call it. But that also is a record of teaching that's super important. And it may be that enabling video production of more and more of these classes-- maybe audio only, but we like video-- so maybe video-- is super important for the future of knowledge.

I don't know. Some of these great faculty members who've taught here, who've taught at other institutions that we've been at or visited-- if only their lectures had been captured. Vladimir Nabokov would be my number one. But that's just me. I think there's something really powerful in the collaboration that we're building with Wikipedia in the sense that learning, in the broadest sense, has this really powerful connection between taking in information, but also activating it, engaging it, doing stuff with it.

And I think it's relatively easier for a program such as MIT OpenCourseWare to go big and deep in the sharing, the instructional pieces of it. It's been relatively harder, though we have tried various models over the years, to create spaces for learners to interactively engage with that stuff in our space. But on the other hand, Wikipedia is just one of those participatory spaces.

And so we have colleagues in open education where the students are having learned some new information. The assignment they're given to activate an experience working with that information is to go make a Wikipedia page better to contribute back to the Commons in this shared collective way. And I think there's something really powerful and essential to be maintaining those spaces for people to work together in an open, sometimes messy way, but in a way that admits everybody in to activate that really high-quality knowledge that they've also taken in.

So for those of us who joined-- those of you who are joining us, for the last 15 minutes, or so I'm Peter Kaufman, associate director of Resource Development at MIT Open Learning, joined by Andrew Lih, a Wikipedian at large, Curt Newton, director of OpenCourseWare. A couple more questions-- one is OpenCourseWare, MIT Open Learning, has 192 people sharing knowledge with the world, sharing MIT's knowledge with the world, working daily on this.

OpenCourseWare is there. We also have new strands of knowledge production, universal learning, which includes universal AI, universal climate. We have the work of the DELTA teams that are doing short-form videos and stuff related to particular physics, mechanical engineering, math. It's remarkable production work that they're doing. There's MITx. There's all of the stuff that MVP, MIT Video Productions-- hats off to those guys who are making this possible tonight-- are producing-- how do we make sure that all of this across Open Learning gets Andrew's tender, loving care and the attention of the Wikipedia community?

That's on you, by the way. But no, seriously, how do we do that? You've been here how long? 8,200-- We're definitely starting in the sphere of OpenCourseWare, as we've known it, and its collection and the way in which we operate because it has that open license connection, which is foundationally important. But as you've noted, there are other modes that MIT Open Learning is creating and sharing knowledge with the world.

And I think many of those have just as much to offer. I know the Wikipedia space, especially those short videos that you've identified-- we've talked about that a bit. And those can not only be extracted from OCW lectures, but also intentionally produced as short-form things. They could be intentionally produced for and licensed in such a way for-- where do you do your thing? Wikipedia, for Wikipedists, for Wikipedia, with Wikipedia in mind.

How amazing if, at the end of every video recording for a class, we were to ask the faculty member to record a 60-second minute about some important formula or some important discovery or some important experiment or some-- It has occurred to me also you mentioned the-- I didn't realize it was 192 people in-- working at MIT Open Learning.

And that's a large number of people. But when we think about our learner community, which is literally millions, many millions of people-- I think we've estimate-- we've reached on the order of 500 million people around the world through OpenCourseWare. And what if one of the things through this collaboration is to also encourage our learners to say, if you've learned something that you carry with you from MIT OpenCourseWare, please do take a look at the Wikipedia pages associated with this?

And if you see something that's-- could use a little love, lean into that to expand our reach beyond just the staff that we've got here and the things that we're producing to empower the whole world. I'll say, it's been quite fascinating working with you folks and realizing when we encounter other folks how influential MIT OCW has been. I keep running into folks in Spain and Brazil and all these other places.

I keep running into folks saying, oh, yeah, we started that based on MIT's OpenCourseWare. I'm like, really? And there are folks that I'm talking about about Wikipedia content. So they wind up lining up in other countries and other language groups all the time in ways I had not realized. And it's fascinating to see that. That's one of my favorite things about getting to do this work-- is to see the way these inspirations play out in different places. We started the first big institution. And I'm just thrilled that year after year after year, we keep losing market share in the sense that the market keeps growing and these new entrants are coming in.

And we continue to do the work. But I know there's so much more to be done. These new models are exploring. It's great. There was a lot of love for-- a huge amount of love for Wikipedia at Wikimania, no surprise. But there was a lot of love that we felt toward MIT and toward MIT OpenCourseWare from the community that just loves knowledge, that loves working with knowledge and volunteering. And thousands of people were there, and online as well. It's a good segue into this question, which is-- this is a story tonight about MIT working with Wikipedia.

But to what degree is this work extensible? Obviously, you've thought about this because you extended your work with the Smithsonian and other institutions where you've spoken at and worked at. But to what degree is this work extensible so that other knowledge institutions can be heartened by what we've done, can learn for what we're-- from what we're doing? Could we create some best practices?

I harp on this in our Zoom calls. But could we create some ways of documenting the work that we've-- how long we been doing it? Not that long, right? It's-- Less than a year. Less than a year. Soon, way less than the 22 years that you've-- but could we do some of this-- you'll be feeling a lot of love these next few days, by the way, because at the start, I mentioned that it's a very busy week at MIT.

We're hosting OE Global 2026. That's Open Education Global Conference. Yeah. And so what? There are hundreds of people coming, hundreds online. These are fast and true old friends as well as its-- you want to say a word about that? Yeah. This is the largest global gathering of people working in the expanding, vibrant field of open education and, as you were just recognizing, people who've taken inspirations and run with it in really fascinating directions.

People are deeply committed to this same shared mission of making sure that knowledge is available for everyone. They're working within the spheres that they have influence. I have a lot of confidence that this opportunity that's opening up that we're thrilled to be working on-- I know we'll get a lot of support. We have a panel, actually, at OE Global. We're doing this again, in other words, in a couple-- Two days.

--in a couple of days. But I think to answer your question, I think we-- more than ever, we're living in this era where people feel like there's-- the label of digital feudalism is more true than ever. We're not buying DVDs. We're just paying rent to multiple entities. It used to be maybe a handful. But now if you want to watch football games, basketball games, you want to watch movies, series, all these things, you might be paying 5, 10 subscriptions a month. And suddenly, that cable bill from yesteryear seems about the same as it is now.

So the great thing about open, about whether it's Wikipedia, whether it's Internet Archive, OCW, OpenStreetMaps, these allies that we have, is we get out of that trap of always having to be a victim of digital feudalism. We get free access to this content. We can do interesting things with it. And I think that's so important in this day and age, to have that as the Commons. What does it continue to mean, and this will keep evolving for-- to hold knowledge as a public good and to keep investing in that?

That's a, I think-- a shared mission that we have a lot of allies in. Yeah, it really is like action. It's what we're up to. We have this Open 2030 Working Group that we started a few years ago. And to put a finer point on the question, that working group branched out to make sure that we could explore how best, for example, to adapt some of the knowledge work that we're doing to learn from in some of the knowledge work that we're doing for and from community colleges, for and from historically Black colleges and universities.

How exciting could it be to make the work that we're doing with Wikipedia, young as it is, and new as it is, galvanize a bunch of-- in other institutions? There are thousands of community colleges. And these HBCUs are-- many of our friends there are embattled. And the support that this could-- how could we inaugurate, I don't know, more systematic collaboration between-- you can't be everywhere, I know.

But-- although you do a pretty good job of it. Is there a way of sharing what we've learned about sharing? I'll leave that for later because I know we've got-- I'm thinking this is the hook that you drop at the very end of the episode to make sure they tune in to episode 2. Yeah, exactly. Well, anything else before we close, because I wanted a couple of thank-yous?

Is there something burning that you need to say to-- No, I just think it just reinforces that we need a federation of folks working for knowledge in the public interest. And this is a good start. Internet Archive, these folks, as well-- we've got to keep growing it. Hear, hear. Amen. So before we close, a couple more thank-yous, shout-outs to some great people beyond our own team-- Phoebe Ayers at MIT Libraries is just the best person in the world to work with.

You're up there, too. And it's Phoebe that we actually thank for introducing all of us sometime ago. Steve Carson and SJ Klein first began to imagine this kind of collaboration between MIT courseware, MIT OpenCourseWare in particular, and Wikipedia almost 19 years ago, actually, 2007. But Phoebe and Curt and Andrew and-- have carried it forward.

And of course, over these years, our supporters at the Hewlett Foundation, the Sloan Foundation, the Arcadia Fund, and now the Templeton Foundation, and others, and a huge, almost overwhelming number of MIT alumni and friends have been so supportive of, even recently, this work, but our work generally, bringing MIT to the world.

So if you're watching, our thanks go out to you for your time today and for your generosity. Thanks to you guys. Thanks to everybody who's been working on this effort with us, some of whom we'll see you at dinner tonight and some of whom we'll see over the next few days at OE Global. And we hope to see our viewers again-- thank you for tuning in-- again soon in person and online.

So thanks again. Thanks.