For today’s blog, we sat down with CEO Dan White to discuss why students need to be prepared to solve open-ended problems and how games like RoboCo can help.
What prompted your interest in this topic?
It really comes down to the fact that outside of school, outside of formal K12 learning, most problems that humans solve in their day-to-day lives, whether that’s at work or even outside, are open-ended problem spaces. There are in many cases infinite potential solutions to any given problem, and that’s a really hard thing for the human brain to navigate. We do a lot better when there are limitations on the options that are available to us. But fortunately – or unfortunately, depending on how you look at it – most problems are not structured that way. And so I think it’s really important that people learn how to navigate open-ended problem spaces. And so I just get excited about anything that helps people on that learning trajectory.
In your Learning Games in 2030 blog series, you predict a shift over the next decade in which educators will begin utilizing learning games to teach “robot-era” skills like problem-solving at least as much – if not more – than core subject areas. Can you offer us a bit more context on this prediction?
The larger context is essentially that the world is changing rapidly. I don’t think that will come as news to anybody. But it’s changing both in ways that we see on a day-to-day basis, and in ways that are entirely invisible to us. In the case of the former, we see for example our smartphones getting smarter and smarter every time. Google Assistant gets better almost every time I use it, for example. But then there’s this invisible revolution happening as well, inside companies that are basically gradually moving towards increased levels of automation, that might be things like customer service, that might be things like part assembly for products, and everything in between. And the really interesting thing, I think, is that a lot of the jobs that are getting automated are not the jobs that you’d expect. There are some really surprising ones in there. The formula is essentially: is there a skill that is repeatable? And is there a skill where the best decisions are made based on large data sets, that humans have a hard time holding in their heads. So those are kind of the things that you’d expect. But then we’re also starting to see machine learning get better at soft skills as well like making music, or writing movie scripts. Now a lot of the work that’s being done right now by AI in that department is sort of funny, because it’s sort of like a kid trying to do something for the first time. But I think that most experts agree at this point that even though we’re at a point where you pick out if an article is written by a bot as opposed to a person, the point at which it’s completely indistinguishable is, I think, so soon. In the blink of an eye we’ll be at the point where we’re interacting with content generated by machines, and we’ll have no idea. In the same way that already deepfakes have gotten good enough that it’s almost- to the untrained eye- it’s almost impossible to distinguish something that’s real imagery versus something that’s faked. So yeah- so I think that the context really is, that this transition is happening really fast, and mostly out of sight, so I think that it’s only natural that it’s going to catch a lot of people by surprise. And I think that the consequences of having our education system caught by surprise in this particular case are pretty extreme. Because we’re talking about whether people have jobs or not. Are we preparing people for an economy that exists or not? Those are the sort of things that we can’t take lightly. And that’s really the broader context for that prediction. In some ways, I’d almost frame it less as a prediction and moreso a concern.
Why is it important that today’s learners are given opportunities to practice and develop these skills?
The stakes are, will students be prepared for the actual economy that they graduate into? Some people think about this in terms of global competitiveness. Are students graduating into the economy in 10, 20 years time going to be able to work the jobs that are available at that time? Versus graduates coming out of other countries? I don’t worry about that so much in particular, as just this kind of broader idea that…because I do think there is a distinct possibility that at a certain point, whether or not you’re employable, will not be sort of a matter of life of death or a matter of, you know, your financial security. I think there is a point at which we will potentially be in a post-poverty condition…but that’s probably still quite a ways out. So in the meantime, the stakes are basically- not only do I have employable skills, but do I have the skills that I will need to add value to the world outside of my academic life? Do I even have the capacity to learn those skills? As we all know, the brain is plastic- everybody has the ability to learn and adapt at any point in their life. But it’s a lot harder when you get older. So when students are at the K12 level, they’re at this perfect place in their intellectual trajectories in order to be molded towards the skills that will ultimately serve them the best. So if we spend that time teaching them to memorize facts and figures, or sort of be declarative content-oriented, or to learn to regurgitate things, then we’ll sort have wired together their neurons in a way that will later in life be really harder to change if we all the sudden ask them to solve large, open-ended, complex problems.
What factors make learning games particularly well-suited as a medium for facilitating these sorts of problem-solving exercises?
So learning games are problem engines, essentially, which means they’re basically little problem spaces. They are, in contrast to what we were talking about before, similar to these open-ended problem spaces that define life outside of school or academia. Learning games are very intentional about creating problem spaces that are not infinite or open-ended. And that’s really important because what we know about learning is that if you’re going to go from point A to point B, we need to give you some handholds along the way. And so learning games sort of simulate, in a way, or give you an opportunity to think in the way that you would need to think in order to solve a large, open-ended problem, but they constrain the problem space enough so that you’re not overwhelmed as a learner, and they give you those first couple handholds. As a learner, you have an opportunity to practice. You’re exposed to an opportunity to interact with these types of problem spaces, without being thrown into the deep end. One of the cool things about learning games as well is that we can constrain that space as much as we want to while still giving the learner a real sense of agency over what they’re doing in the game space, and good games do this all the time, even if the actual amount of agency that the player has is relatively low. Games do this really great job of making you feel like, say if you’re playing an open world game, the game’s really good at making you feel like you can go everywhere and do anything. But in fact, the actual verbs or actions at your disposal are quite limited compared to the real-world. So even take a game like The Elder Scrolls V: Skyrim, where you feel like you can travel great distances, you can make choices as you see fit, you can talk to anyone you want to, you can fight anything that you want to…but when you really boil it down, versus the agency that you have in the real world, it’s a pretty constrained set of verbs. You can fight, travel, craft, customize your clothes or armor, or ride a horse- but you can’t make art. You can’t play an instrument. You can’t hold a parade. You know- we sort of take it for granted, but there are infinite things you can do at any point of time in the real world, but in these constrained game spaces- even the ones that are designed to give you the impression that you have complete agency and control- you’re actually quite limited in what you can do. And that’s a great environment in which to prepare somebody to think about an open-ended problem space, but still with training wheels on.
What are some examples of existing games – commercial or educational – that you feel offer compelling open-ended problem-solving spaces?
I would be remiss if I didn’t mention our upcoming game RoboCo. When we started Filament Games about 16 years ago, the project that kicked off the company also happened to be my master’s thesis. And the game was an ocean science game in which you play an ocean science researcher on an alien planet, trying to understand what the impacts of a downed phosphorus transport are on this underwater ecosystem. And we very intentionally created a problem space in which we give you the tools and science and the conceptual understandings of how scientists collect data and use that data to form arguments. But beyond that, it’s really up to you as a player to decide what data you want to collect, and what you want to do with that data. So the overall goal was to sort of give you the impression of science as a practitioner operating on the cutting-edge of your field, as opposed to how it’s often taught, which involves learning about the things that other people in the past discovered. Across the last 16 years we haven’t had a lot of opportunities to create games that were that open-ended, or afforded the player that much agency until RoboCo. RoboCo has allowed for that kind of design, and in many ways is a really cool project for Filament because it harkens back to our roots of giving the player a lot of creative agency, and a lot of control over their own destiny within the problem space. In the case of RoboCo, you’re not decoding an ecosystem, you’re building robots, building bespoke robots in order to solve a particular problem or challenge, whether that’s delivering a sandwich or crossing a gap to turn a valve or cutting branches off a tree. At the end of the day, the two experiences are super similar in that the player has a ton of bandwidth to make decisions, and I think importantly, interesting decisions that have real ramifications inside the game space about how they go about solving that objective. One of the litmus tests for determining whether or not you’ve crossed over this critical threshold of player agency is the question of – if I give this game experience to 100 different learners, how many of them are going to solve the problem in the exact same way. If the answer is less than a dozen, then you know you’ve created a space that really affords the player a lot of agency.
Our upcoming robotics game RoboCo is a largely open-ended sandbox experience – what strategies have you and the team embraced to encourage players to flex their engineering ingenuity?
The objectives themselves encourage this – so this may sound slightly in direct contrast to what I just said, but to the extent that it is contradictory, it is intentionally so. Each challenge level is engineered such that the objective of that challenge is designed to explicitly force the player to think differently about how they design their robot solution. So, for example, in one challenge, the rules of the space and the way the space is designed is set up to really encourage the design and development of some kind of catapult-like object, versus another one really might be designed to encourage the use of a part like a piston, for example. Another one might be designed to encourage you to think about leverage, because you have to build a robot that’s stable- even when it extends itself out in some way. Another one might require you to think about how your robot will interact with gravity as it tries to reach an objective that’s really high up. The core premise that you have complete control over the solution is true across each of these contexts, but each context is specifically designed to force you outside of whatever comfort zone you’ve established with the previous solutions that you engineered. I think that’s super important, because that’s one of the other natures of open-ended problem solving, and that is that the solution that worked last time is probably not going to work next time. You always have to be innovating.
Bonus question – any other ideas for open-ended problem-solving games?
This would be a weird game, but I actually think it’d be really useful. One of the biggest open-ended problems that almost everyone has to solve in their lives, is that they have to get a job at some point. This is one among many examples of things that everybody has to do at some point in their life that school for whatever reason is like, “this isn’t an important skill to prepare you for.” It’s sort of like finances, or relationships, or politics- all things that every human has to do in their life to survive, that school kind of glosses over. While schools are like, “no, we really need to focus on Algebra, solving for X is definitely more important to your ability to find a job.” When you break it down, like any specialized domain- there’s specialized terminology, and specialized skills, and conceptual understandings and learning objectives, just like anything else, and that extends from how you craft a good resume, to how you interview well, to even the non-obvious things that a lot of people don’t even do- good followup, for example. Or networking leading up to an interview. Or informational interviews, things of that nature. Getting a job- particularly a competitive job- can be, I think, as complex as a really complex strategy game. It’s an interesting problem space because there’s not a right or wrong way to do any of it. There’s certainly best practices, but there’s not a standard way that you should always do it. One of the other things that makes it an interesting problem space is that you’re competing against a bunch of other, presumably intelligent people, and you want to stand apart from the crowd. In some ways, part of what makes you a successful job applicant is often thinking outside the box, and doing things differently than anybody else. I think there’s a compelling game there. I think it could even be something that people play on Steam for fun, given the weird things that people play on Steam for fun. The agriculture idea is basically just- oftentimes, when we’re developing learning games, we have to stretch more or less to convert the content into a learning game format. Some content just naturally lends itself better to learning games- and agriculture I think is one of those usually underserved areas in terms of education that also happens to be, I think, perfectly suited for game-based learning because it’s very much- it’s really this unique combination of hands-on practice, like there are certain skills like- “how do I plow a field?” kind of skills. But it’s also very much a strategy space as well because you have to make decisions about timing, and risk management, and finances. I mean, it’s essentially running a small business if you’re participating in the farm. It’s this really cool combination of strategy-level thinking, and then just straight up procedural knowledge about how you actually accomplish tasks in a farm environment. It reminds me of a game that probably not many people have played called Wing Commander: Armada. This was one of the first games that really let you command your flagship, you control your capital ships at a strategic view, and then you jump into the cockpit and actually fight the battles. There’s a lot of games like this now, but this was one of the early ones that did it. There’s something really satisfying about jumping back and forth between the boots on the ground experience, and then the strategy-level experience. Agriculture is one of those disciplines that uniquely lends itself to that dichotomy.
Anything else you’d like to add?
I think that people are not talking enough about how the world is changing very quickly, and at least from where we sit, I’m not seeing schools adapt fast enough. I don’t think it’s hyperbolic to call it a crisis. I think there’s going to be people that graduate into a workforce that they’re essentially entirely unprepared for, or ill-equipped for. A lot of people are going to learn the vast majority of the skills they need to have after K12 school, which is just ridiculous. But that’s the direction we’re headed right now. If K12 wants to stay relevant right now, they have to acknowledge that grades don’t matter, test scores don’t matter, it’s very hard to design a test (particularly a multiple choice test) that can in any way be associated with the skills and abilities that will need on order to work the types of jobs that are not assumed by machines.
Read on for more educational game industry insights: