What if you knew what each person in your organization really knew? What if you could know what they were capable of, that might surprise you? What if they knew what they were capable of? What if your organization was filled with superheroes and heroines who hadn’t revealed their secret identities just yet? What if you could team each of those “superheroes in waiting” with the best mentor, coach, guide, and leader in the world to help them find and take action on their capabilities; level-up, when the chips are down. Imagine a future where you could do this, not for a handful of your high potentials but at scale, across your organization. Imagine you could dynamically play out 10s, 100s, 1,000s of “fantasy learner action league” combinations to explore where the highest value investments in learning will lead to outsized returns. Consider when each person is able to spawn hundreds or thousands of “digital stunt doubles” that let them play out their own futures to find the right pathways ahead. What if those investments could be linear in time and cost, but return exponential value the more the world around you accelerates into the unknown.
We need to act now
Unesco estimates that 470 million new students will demand higher education in the next 15 years. will we really build eight universities a week from now until then to meet that demand or could we take a page from the way other industries scale yet provide a personalized touch?
Meet Danny Lange. He is the head of AI at Unity, they make authoring tools to design, build, and deploy video games. They reach over 700 million people around the world, and have built a platform that enables new experiments in play, and more recently simulation, that can be run by, at the current count, a community of over 750,000 developers. As the former head of AI at Uber and Amazon’s cloud, Danny saw the power of using instrumentation—basically virtual and real sensors—on everything from web pages, and cell phones, to cars and roadways. By detecting exactly what people really did from moment to moment, how they learned, and how to teach an algorithm how to play well with their human teammates, he’s been able to deliver orders of magnitude larger impact.
Currently, he’s helping automotive companies build virtual world “avatars” that can drive more miles than all the humans on the planet could do in a thousand years. They are simulating not only the physics of the vehicles interacting with each other, but how gravity, friction, inertia, road conditions, and even daydreaming pedestrians all interact with each other. The companies now build massive “synthetic learning” datasets that would otherwise have been impossible—and dangerous—to capture by driving human or autonomous vehicles around. In the real world, those same automotive companies would have to sensor up and drive vehicles from now until the end of the next millennia to get the insights they now discover in weeks or months.
Before the rise of cloud platforms and commoditized sensors, before the rise of world-spanning networks of gamers joining tournaments, building fantasy leagues, and going from newbies to legends, before the rise of machine learning as a service, people played games. Play is not only a critical component of learning, or as Mr. Roger’s once said, “play is really the work of childhood” but play creates what neuroscientists and behavioral economists call “desirable difficulty” or “productive struggle.” Play is in context, creates psychological safety, allows for learners to become attuned to cues to take action, and supports response rehearsal so they aren’t caught like a deer in the headlights when a new reaction is necessary. Play enables learners to develop skills and habits through trial and error, has been shown to reduce “decision myopia” and delivers immediate and variable rewards while also building a sense of intrinsic motivation at the individual and team level. Play is critical if you want to accelerate learning, move motivation to habit, and build resiliency in an increasingly volatile future.
So what does this have to do with learning and development on the job?
With the rise of our ability to instrument actual work at scale, sense what people really do rather than what a resume or survey says they can do, and the ability to enlist tireless machine learning teammates and simulations that have more insights than any single learner, team, or leader could ever gain in one lifetime, we stand at the threshold of a new era in learning and development. If we want to enable our workforce to grow their capacity, prepare for their next role, build their literacy and fluency in leadership, followership, and the skills once thought of as soft but now considered to be power skills, it’s time we built learning directly into their workflows and shifted from measuring if someone took a class or got a certification to uncovering what our workforce already knows but can’t demonstrate in classic proxies, and enable them to discover how they get from where they are today, to where they can make the biggest impact tomorrow.
Early signals from the future
Mixed or Augmented Reality (MR/AR) is not quite ready for primetime for many tasks but we expect that to change within the next few years. Consider all the aspects of context that the current generation of systems can detect and imagine what you might be able to do with that information to set levels of competence and confidence in your workforce. An MR system can not only sense the surroundings—using dynamic 3D scanning, microphones, gyroscopes, and GPS—but can also detect where the learner is looking (or what the learner is ignoring at their peril), their heart rate and galvanic skin response (are they calm or nervous, doing something for the first time or are they old hands at the task), the speed and fluidity of their action, and capture how those activities change over time. Instead of having weekly, quarterly, or yearly performance reviews as feedback, which can become disconnected from the person’s actions and muddy the waters around the actual activities that lead to them being their most effective, teamdmates can learn in real-time, in-context, and take ownership of their own mastery.
Using just a smartwatch and a few cameras Autodesk Research created a “virtual AI-foreman” that directed a stream of random humans conference-goers and four robotic teammates in a dance that started with a pile of bamboo, some spools of string, and various connectors that by the end had built a new large scale living structure called “the hive.” Could we have AI-coaches that help us individually and in teams stretch and grow over our careers to build something out of all the skills, mindsets, and toolsets we have at our disposal?
Sensing and Context
We can find early signals from industries that are adopting these tools today, for instance BAE systems uses augmented reality when their employees are in real-world situations to deliver a 10x reduction in training time by giving instant feedback, in context, via visual overlays as employees learn tasks and build mastery. They’ve also seen a 30% increase in the effectiveness of their training and the impact is directly measurable. They’ve seen upwards of a 50% reduction in manufacturing assembly time.
Machine Learners Never Sleep
Bam Ireland has been around for over 50 years and is the largest civil engineering contractor in that country. For civil engineering the risk to their employees, subcontractors, and their profit margins are are all directly related to how effectively they shape their contracts, specifications, and drawings and how those directives are converted into a completed bridge, wastewater treatment facility, roadway, or piece of critical infrastructure. In the process of construction, a mistake in the design phase that makes it to the site during construction can cost them hundreds of thousands of dollars, a safety incident can not only cost time but also risk human lives. To tackle these challenges and build a more effective workforce they used a combination of data from their project management system, mobile devices on-site, and a collection of machine learning algorithms on a platform called “Construction IQ”. They saw a 20% improvement in safety and quality on-site. Back in the office they saw a 25% improvement in decision-making, and focus on high-risk issues. A secondary benefit is that they were able to improve their accuracy and predictions about how their subcontractors and employees would perform on future projects. The lessons within a project ripple outward and build more value across projects and deliver increasing benefit as the information pool grows and the algorithms get tuned within a loop of human and machine learners.
Caught in the In-betweens
Confidence comes from knowing what we know how to do, and having a sense of agency about actions we can take that will lead to impact. These peak moments also lead us to wonder what we might do next. How we might step into a new role and move upward into new adventures.
But as we move downward from a healthy sense of confidence and success, we go through feelings of uncertainty, struggle, and at times apathy. We disconnect, isolate, try to rally, lose confidence, face self-doubt, and when we fall too far and hit rock-bottom we can become depressed and want to give up.
Consider this spectrum along a vertical axis and realize that each and every one of your teammates is somewhere on that spectrum at any given time. Over the course of their lifelong journey, they will spend times facing mid-career moments stuck in the “in-betweens.”
By building learning into your team’s workflows you can look for ways to capture and guide them towards shared personal and workplace goals, detect the strategies that work best to help them navigate these uncertain periods and convert them into hero or heroine journeys that build their capabilities, foster fruitful connections and encourage upward mobility.
These journeys can have micro-mentoring moments to resolve precise challenges. They can happen across teams within your organization in powerful and privacy-preserving ways. they can help people that are otherwise left behind find ways to be stars through alternative routes. The journey could last a lifetime and span age ranges from when the learner is young thru when they’ve retired and are looking for their next chapter and reached a place called “the blue zone.”
Ways to get started?
Consider ways of mapping your employees’ opportunities for learning both at critical “in-between” moments in the journey, and as they move through their careers. What learning spaces, tool flows, and mindsets do you want to factor in or foster? What network and community effects do you want to shape? How might you sense and automate key moments of learning and change? What policies and incentives do you want to put into effect and how will you learn their effectiveness and impact within your org? How might you use simulation, competition, and co-opitition to drive rapid upward advancement and team play? What events, programs, environmental interventions and other wildcards will you use to get started and spin up the flywheel?
Emerging platforms hint at what might be possible. For instance Kaplan performance academy has a competence/confidence mapping tool combined with an AI-driven engine to foster organizational growth and signal when someone has low confidence but actually has a high level of competence or when the reverse happens and someone is overly confident about a given role or task but is going to either struggle from the illusion of mastery or fail and miss the mark. There is new research into mitigating this illusion and accelerating learning via “tools for thought” that team human learners up with algorithms that take advantage of the neuroscience of memory, and attention.
Autodesk’s Learning Engine is exploring aspects of goal-directed learning within their platform workflow by sensing and visualizing a users’ level of mastery within the tool and tying it directly to peer-generated tutorial videos as well as community suggested insights on which capabilities the user should learn next to guide them to on their learning journey most effectively.
We predict that in the next three to five years these platforms will allow your teams to make linear investments in learning that have exponential returns in retention, confidence, and the ability to take actions and synthesize ideas in novel ways to drive more daily creative problem-solving.
We need to have a discussion about neurorights.
All of these benefits come with costs as well. If we don’t consider how we’ll protect the most personal data sets, those that tell us how we and our teammates think, process inputs and outputs, and generate responses, we’ll be in trouble. The emerging discussion around neurorights needs to happen now, not when we have chips to put in our heads.
Next up? Data governance, neurorights, and the ethical dilemmas ahead.