Back to Blog

The Robots Are Coming

Filament Games | Educational Game Developer

Kai-Fu Lee, author of AI Superpowers, recently predicted that “within the next 15 years, 40-50% of routine jobs in the US will be replaced by artificial intelligence.” In other words, by the time today’s kindergartners graduate from high school, employers will care little about automatable competencies.

As the founder of a fourteen-year-old company, I have a poignant awareness of how quickly 15 years can pass, which is why I find it unnerving that conversations about teaching and assessing uniquely human skills are still nascent. As Ken Kay, founder of the Partnership for 21st Century Learning, recently told EdSurge, “Leaders at the forefront of the 21st century learning movement tell me they still see too many students sitting passively while teachers deliver instruction; too much technology is still used to replace routine tasks rather than turbo-charge the experience of learning.”

Indeed, we can’t teach 21st century skills with 20th century pedagogies. To prepare students for the imminent robo-workforce, we need to teach them how to be really good at being humans, and that means teaching skills that are difficult to automate: problem solving, creativity, collaboration…the usual suspects; often referred to as 21st century skills, higher-order thinking skills, or my personal favorite, future-facing skills.

Easier said than done, of course. Convincing schools to prioritize these skills will be difficult so long as administrators are held accountable to high-stakes tests that measure traditional skills. K12 assessments are the quintessential example of the tail that wags the dog. The curriculum mirrors the test, so if the tests evolve to incorporate future-facing skills, the curricula will likely follow suit. Of course, the instrumentation must evolve as well, because multiple choice questions are a poor way to assess soft skills.

Enter video games.

When intentionally designed for assessment, video games can be ideal tools for rendering a subject’s soft skills visible to an external observer. A video game designed to assess collaboration, for instance, can immerse the subject in a rich problem space that solicits collaboration-oriented solutions at every step. The game can position the subject at the center of a multivariate matrix of decision points, allowing them to generate unique solutions that yield fingerprint-like results (i.e. Jeff’s collaboration style emphasizes setting expectations early and building consensus longitudinally) rather than numerical scores (i.e. Jeff’s collaboration skill is 80 out of 100) or, worse yet, dichotomous verdicts (i.e. Jeff is bad at collaboration). In other words, well-designed assessment games can yield insightful/actionable stories about people; not just summative declarations. And unlike traditional assessments, games can capture high-fidelity streams of gameplay data, allowing external observers to view process data; not just outcome data. Remember how your grade school math teacher always wanted you to show your work and not just your answer? Same thing. Every action Jeff performs in the process of navigating our hypothetical collaboration-oriented problem space can generate a bread crumb that offers a window into his cognition. Cumulatively, Jeff’s bread crumbs likely tell a more interesting story about his skill at collaboration than the result he achieves at the end of the game.

Several startup companies are already working to harness these powers. Imbellus builds simulation-based cognitive assessments that evaluate “how people think, not just what they know,” and they ultimately endeavor to replace standardized tests like the SAT, ACT, and AP exams. Another company, Scoutible uses mobile games and machine learning to measure cognitive and personality traits like leadership and creativity. Both companies are still in their infancy and are exclusively using their software to help corporations assess prospective hires. Nonetheless, if they prove that games can effectively assess soft skills, similar technologies will inevitably show up in K12.

The current limitations of standardized tests and the resulting vestigial nature of school curricula are, I believe, the number one reason why games have struggled to be adopted into schools. So long as the goal is to improve student outcomes on multiple choice tests, superficial drill-and-practice pedagogies will reign supreme over game-based learning, project-based learning, and other progressive pedagogies. Once assessments evolve and are capable of measuring uniquely human skills, however, games and simulations will proliferate throughout the K12 ecosystem.

More insights from CEO Dan White: