GEORGE SIEMENS: So if you were to look at the field today, just sort of where it's gone over the last 20 years, where it sits now with open education shifting, more so from just content to the mappings of that content to outcomes, to the use of learning analytics to do a better job of producing or creating an impact for students and student success, and to evaluate just the cognitive development, the institutional influence, and so on of analytics, where would you say your primary concerns are? Even if you were to articulate my top three concerns for the open education movement going forward are--

DAVID WILEY: Oh, so you've already mentioned the way that publishers have kind of ceded the battle to try to win the war. So the war is really going to be on platform, not on content. And content was an easy battle for the community to fight, because anybody with a PhD in the discipline could sit down with a word processor and create something that had a hope of being competitive with what other textbooks somebody else might be using.

When you talk about trying to replicate platforms for personalization analytics and things like that now, the team of people in the open community who can really compete head to head with what the publishers are doing narrows drastically. And I think in as much as these assessments and competency statements and content, if those are all open, I think it makes total sense for them to be placed into a variety of platforms to see, how can this platform's special sauce make it more effective, or give a different perspective or a view on it? But at some point, it seems like the content, the assessments, and the competency statements start to become less and less interesting as they're kind of subordinated to the platforms. And if all the platforms are proprietary, that's a problem.

GEORGE SIEMENS: Well, and you consider too the issue with networks that we don't hear about often enough, is that networks don't produce many winners. On a given day, we look at early on, I remember mid-2000 when social media wasn't called that, at that point we called it web 2.0. But there was a mess of sites. You'd be on Ning one day, and then the next day you would be on--

DAVID WILEY: Friendster.

GEORGE SIEMENS: Friendster, early stages of Facebook. I remember at University of Manitoba, we had built our own social network for students to interact. And so you worked with dozens of platforms. And yet today if you want to connect to reach out to people, you're on Twitter, you're on Facebook, you're on Instagram, you might be on Pinterest, and you might be on Snapchat if you figure out how it works.

And that's it. So that's the issue. And I think publishers, we're going to see some similar network effects where the platform owners for adapted and personalized learning, unless there is a concerted attempt to create a vision, and to create an open landscape where that's possible, we're going to see sort of a Facebookization of open education as well, which is in the long run, why create for x number of platforms when 98% of your users are here?

Or just consider mobile phones now. You're designing for Android, you're designing for Apple. Who else are you designing for? You captured the majority of your audience there.

And so I think when I'm concerned, when I look at the future, I'm concerned exactly that is that the platform effect, which gives us a lot of really good benefits. It's so easy to log into Facebook and see what everybody's up to. You join a new service, you just use your Facebook or your Google ID to go onto the service. It's easy. And easy sometimes is a bit of an enemy for openness, because it's easy for me as a faculty member, if I'm not paying for that textbook, to just use a textbook that gives me the PowerPoint resources and away you go. So the platform issue is a significant one to look at. Is there anything else you're seeing though, that you find alarming or disconcerting, or even just areas where you think your main focus will need to be going forward?

DAVID WILEY: You know, I'm so focused on the platform problem right now. It's almost kind of hard to see past it, because if you lose on platform, it's hard to see how you can win on anything else.

GEORGE SIEMENS: So can you explain how Lumen is a platform problem? You mentioned this earlier in one of our discussions. But now that we've had a few weeks under our belt with the content and we're moving toward wrapup, I think hopefully course participants will see why that platform view is such a concern for you. But can you talk a little bit more about how you're addressing the platform problem to your current projects, namely Lumen?

DAVID WILEY: Yeah. So the first, most obvious answer is that the overwhelming majority of the platform is open source. It's a GitHub, you can grab it and check it out and play with it. And you could stand up a Lumen competitor tomorrow, between the open content and the open pieces of the platform. There are pieces of the platform that aren't open yet, but I think watch that space.

On the black box algorithms side, I think there are some-- I don't know if they're useful, but I think at least interesting attempts that we've made there that for example, in dashboards presented to a faculty member, this is George here, here are the five students in your course that need the most attention right now. Reach out to them. You'll see that our approach is very, very light on visualizations of graphs and charts and things like that. And it's very heavy on expository prose that explains to you, when it says, here are five students who are trying but struggling. Then there will be a short paragraph beneath that that in English language describes what algorithm, and how the algorithm functions that made people land on that page.

So for example, these students all have done the reading. But they failed their first attempt at the quiz. And they're not engaging with the self-checks and the other opportunities that are available. So the way that that gets implemented in software is described in plain language for faculty members so that even if you don't have the technical capability to go in to audit the code itself, you do have some understanding of, here's why people are showing up here.

GEORGE SIEMENS: So at what level do we start to say, you can't be an educator today if you're not a data scientist, lite or a programmer lite person? Are we at that stage where we're discussing here the fact that the content which is traditionally our practice as educators has been about developing people. We use content as a primary development mechanism. There was a knowledge component, but there's also a mentoring and a beingness to that experience. Are we now at a stage though, where some level of meta-cognition about the analytics, and about the content that's being used, and how models are being created for students where that needs to be a part of our education practice in any course, much like in the past, let's say critical thinking has been embedded in all aspects of the curriculum?

DAVID WILEY: Well, that's certainly a problem that we think a lot about. And we think about it even a little more expansively than that in the context of our analytics work, because again, our analytics work is mostly forward facing to the kind of faculty member that will end up at a community college, that'll end up teaching a 5-5 or a 6-6. And it comes from a particular academic background that's not the same background you'd have if you're at MIT or if you're at Stanford or someplace like that.

And we really do think about it very much as this faculty member has no background in the learning science. They have no background in instructional design. They have no background in data science. And they have no background in the behavioral economics of it either.

And that in some ways, or in many ways, is the most interesting piece, because you can understand what needs to happen. You can understand the learning science piece of it. You can understand what kind of tools and pedagogies and assignments and technologies might bring to bear to encourage that to happen from instructional design perspective.

You can see all the data and try to get inferences about what the data are telling you, about how that is or isn't working from the data science side. But at the end of the day, it turns out that you provide so many opportunities to students that they just choose not to engage in. But you know that they're there, they're paying money, you know that they want to learn. They want to be earning decent grades, and they're not engaging in behaviors that are clearly in their own self-interest. So it turns out there's a behavioral economics piece of it as well, where there's some nudging or there's something that needs to go on there.

So you look at all the things you have to understand, and we kind of have a bulleted list of faculty that are not learning scientists, they're not instructional designers, they're not data scientists, they're not behavioral economists. The job of somebody who wants to make that easy for them is to bundle all that up in a way that in kind of one paragraph, you can explain in natural language what's happening.

Of course, it's not the full richness of all the understanding that you would have had if you went in to get a master's degree in data science or something, but enough to let you leverage tools in powerful ways to promote and support better learning. The overwhelming majority of faculty right now don't have the first-- they haven't had a single semester course in just the pedagogy of their own discipline, or anything.

They did a PhD in physics and they got a job teaching physics. And you'll teach the same way that you were taught, and that's kind of the assumption, right? The idea that if we're not currently training them just on the basics of teaching, that we're somehow going to succeed in also training them on data science and the instructional design and the other pieces of this, it seems kind of far fetched to me. I think our best hope is embodying those things and tools in ways that are transparent and usable and open.