I met William Zhou (entrepreneur, CEO, Forbes 30 under 30) in May in between braindates at C2 Montréal and we had a good conversation around education, his company Chalk.com, intelligence amplification, machine learning, centaur chess and humans learning from AlphaGo. I thought it would be interesting to take up some of those same topics in an interview here.
Patrick Tanguay—You’ve said before that you have a passion for education and are constantly learning about the industry, do you find that gives you an edge vs entrepreneurs who might simply aim to disrupt?
William Zhou (WZ)—Entrepreneurship itself is about learning, experimenting, and doing. As an entrepreneur, you have to be in a position where you’re growing faster than the company you’re building. Otherwise, you’ll squander the opportunity.
Many education technology companies are started by founders with backgrounds in technology without much prior understanding of pedagogy. To make a real impact, these founders have to understand the methods of teaching and learning. Education entrepreneurs have to gain mastery of teaching without being a teacher, and that’s no easy task. Without such knowledge, the industry will be littered with beautiful applications that do not actually move the needle in student outcomes.
At Chalk.com, we like to think of ourselves as a learning company rather than a tech company. Understanding pedagogy is the responsibility of everyone in our company.
The education system is largely based on a factory model, “chain producing” students for the workplace, often years behind current job markets, never mind in a few years when today’s students start working. What are your thoughts on that and does Chalk address the situation somehow?
WZ—The factory model was designed to transfer knowledge as efficiently as possible. It certainly increased literacy amongst people, but missed out on fundamental skills. This model of education also created uniformity, leading to an epidemic of apathy towards learning.
A high quality education cannot be automated or mass produced.
What we need is a model that doesn’t rely on rote memorization, but rather, we should be teaching fundamental skills. The system needs to accommodate your needs to give you the best chance to succeed. A high-quality education is one that captures the imagination, engages the learner, and builds on skills.
A high quality education cannot be automated or mass produced; it must be personalized by a professional who understands the readiness, interests, and preferences of each learner. Why? Because learning is innately human. The problem is that this has not been possible at the K-12 level within the current system. We would all be better off with one to one instruction, but you simply can’t have 30 teachers for 30 students. As a society, we simply don’t have the resources for it.
Technology can help education in two main ways. The first is that teachers are overwhelmed with work. When used right, technology can help teachers save time. Technology also dramatically reduces the friction required for teachers to collaborate over large distances. This is the case with teachers collaborating over lesson creation online. This transfer of knowledge and techniques is especially helpful for newer teachers.
The second way is by giving educators insight into their classrooms and their students’ needs. Personalizing an education for every student in a 30 student classroom is impractical, teachers don’t have that much time. Technology can give insights to teachers, allowing them to cohort their students into 5-6 smaller groups that have common interests. Teachers are then able to differentiate their instruction to these groups. It might not be one to one instruction, but it is personalization.
Curiosity, a love of learning, agility and a growth mentality are often mentioned as skills to “future proof” students. How do you think this can be taught in school and do you have some other skills or characteristics you feel should be part of a student’s toolkit?
WZ—Rarely do we see children coming home yelling they love school. Often, it’s quite the opposite. At some point, we have to start wondering why. Sir Ken Robinson famously said in one of his TED talks that creativity is educated out of us. Why is it such an unpleasant place? Clearly, students aren’t being engaged in classes. Perhaps, it has to do with how material is passed down, or the rote memorization of facts that can easily be found via a Google search.
When we look at what makes us love our job. It’s often teamwork, project, or purpose. Imagine if your job was to memorize some mindless information and not know how it was actually relevant to you. You would hate your job. Well, that’s exactly what we do to students. We love to learn when we see the purpose. As a whole we should be doing a better job at engaging students. For the teacher, it may be to personalize more, to create project based learning environments, or to explain why we’re learning a particular subject. For the parents, it’s about being there to support their children and being part of the learning process. Once children see the purpose and find the interests, they’ll begin developing these necessary skills themselves. Ultimately, students have to take charge of their own learning.
A message we often see coming from Silicon Valley is that technology, automation and algorithms destroy jobs. You see technology as more of a tool to enhance teachers’ (or workers in general) capabilities. Can you expand on that?
WZ—The buzz in education is all about personalization. People think that perhaps we can automate away the teachers. Personalization has been identified as the vital next step for education, enabling the next generation of students to develop the skills needed for the 21st century. However, this is not a new concept. In fact, this idea is almost a century old.
In 1926, Sidney Pressey invented a machine that had two modes of operation: teach and test. After reading through material in the teach mode, a student flicked the control to the test mode and proceeded to answer by pulling down one of four response keys. To give the illusion of progress, the machine scored the response and recorded the number of correct answers. A reward dial could be added so that when a student got a certain number of correct responses, a piece of candy would drop into a small dish for the student. Students worked at their own pace and learning was individualized. This invention was called the teaching machine.
High-quality learning environments are deeply creative, human oriented, inquiry based, and emotional
Unfortunately, teaching machines do not build more resilient, creative, entrepreneurial or empathetic students. Nor do they build on skills like critical thinking or creativity. Adaptive learning systems are reductionist in nature and primarily focus on those things that can be easily digitized and tested. They fail to recognize that high-quality learning environments are deeply creative, human oriented, inquiry based, and emotional.
Instead of replacing teachers, I view technology more as an aid to teachers, especially when there are human components. Too many people are hung up on artificial intelligence. Instead, for many industries, we should be looking at intelligence amplification (IA) to create immediate impact in decade old industries.
This is also the mission we laid out for Chalk.com—to support professionals in bringing personalization to education. We see Chalk.com not so much as a tool as it is a personal assistant for every teacher.
Education is a very personal and human endeavour. When you couple great teachers with enabling technologies using methodologies like blended learning, we finally have the tools and methods to achieve personalization at scale. When we do this, we’re not only bringing quality education to all, but also helping give everyone an equal opportunity to participate in the future.
It’s a completely different type of learning and teaching but I’m curious to hear what kind of reflections you might be having around machine (and deep) learning, AI and teaching neural networks?
WZ—We’re excited about the developments in machine learning. Data driven instruction will enable an entirely new way of teaching. Right now it’s very difficult for educators to keep track of students who are being left behind. But leveraging machine learning and data, we can take a lot of that weight off of educators’ shoulders, we can determine which students are struggling with the concept, and which are excelling at it. We can keep track of the learning styles and interests of the students to enable the delivery of not one lesson for everyone, but instead a couple of lessons tailored to different groups of students so that they can learn at a different pace. This will help students become more engaged with what they’re learning, and ultimately help them to do better in school.
Before we get there, we have to gather the right data first. We also have to do so without adding work to a teacher’s day. Again, all of this has to aid the current human interaction, without making a drastic change to how the actual learning happens. The wonderful thing about what we do at Chalk.com is that we can collect this information without adding more work to the teacher’s day. Teachers are entering data into Chalk.com as part of their daily workflow, and with that information we’re able to help them with their differentiation of instruction.
Data collection, algorithms, machine learning, etc. All areas where we’ve seen (warranted) pushback having to do with privacy and bias. It’s hard to keep data private and it’s analysis and usage free of the programmers’ biases (even “good” ones). Same goes for the data and choices made in programming algorithms. Particularly interesting to think of those issues in a field dealing with minors. Has this been part of your thinking so far?
WZ—Companies have to be mindful when dealing with sensitive student information. That is not to say that such analysis can’t be done anonymously to learn collectively. We’re starting to see various privacy pledges sponsored by companies, mainly to promise never to advertise or sell such data. This is a good start. We can draw a lot of similarities to the health care industry also. There certainly needs to be guidelines and perhaps even regulation. The last thing we want to do is to profit from selling to students and selling their data.
William also covered some of the same terrain when he gave this talk at TEDxKitchenerED.
Header image by Pavan Trikutam on Unsplash.