KnowAtom Interview with iRobot CEO Colin Angle

Colin Angle, CEO, iRobot. Photo from iRobot's website.Today I am proud to share a fascinating interview with iRobot CEO Colin Angle about artificial intelligence and its contribution to effective, engaging STEM education.

FRANCIS VIGEANT, CEO, KnowAtom: We have a well-defined cultural idea of what a robot is from shows like The Jetsons or Lost in Space. Can you give us a more accurate definition of what a robot is today?

COLIN ANGLE, CEO, iRobot: It's true that when authors and filmmakers originally started writing about robots, they essentially envisioned machines that can do what people do. Mechanical people. Today, in order to make a practical robot, it's very rare that building something in the shape of a person is the right solution. It's just too expensive. Legs and arms, for instance, are very expensive to build, and unnecessary to the definition: People feel very comfortable calling the vacuum cleaner Roomba a robot.

In essence, a robot is a sophisticated machine that has sensors to perceive the environment within which it operates, some type of brain that thinks about what it sees, and motors that turn that information into physical action. There's something more to it, though. If you think about it, a modern car fits this definition in that it can perceive its environment and taking action based on what it perceives, yet most people don't think of a car as a robot. The essence of a robot, I think, is that in some sense we think of it as being alive. This may be in a very limited fashion, but we do think of robots as having feelings and emotions that you wouldn't ascribe to a blender. People may name their cars, but I still don't think they think of those cars as somewhat alive in the same way they do a robot.

VIGEANT: Speaking of naming, many people name their Roombas. Do you have a percentage on how many people do this?

ANGLE: Nearly 90 percent. Either people will give it a name or they will call it "Roomba," as though that was actually its name rather than a product. The amusing thing is that if you don't own a Roomba and you're thinking about buying a vacuum cleaner, and you're asked whether you would name that vacuum, the answer is an almost aggressively violent no. People are indignant about the very idea. Yet when you bring the robot vacuum home, turn it on, watch some television, look over and see it working hard on your behalf to clean your home, this instills in people a sort of affection. They think, "It's part of the gang. It deserves a name." And so it becomes Rosie or Roswell or what have you. It's very, very common. This points to us thinking of the robot as in some way alive.


Roomba [photo from iRobot's website]

VIGEANT: So there's definitely a relationship forming with the machine. I know you've had significant experience with artificial intelligence. Do you see a difference between the way robots think and the way humans think?

ANGLE: There is certainly a difference today. It's very difficult for a robot to understand metaconcepts and philosophies, for instance, so we certainly cannot ascribe anything like human intelligence to a robot today. Robots have much greater challenges in attempting to learn. I mean, the human brain is remarkable as a learning machine, and is much more advanced than anything we can build today.

But robots are starting to have some traits that could be viewed as emotions. I believe people have emotions because they allow us to resolve and act definitively when there's not enough information about the world around us to make decisions based on pure logical reasoning. That same ability is useful to a robot. Is there something magical there? We'll see. While artificial intelligence has been embodied in robotics to date, though, we're laughably far from human intelligence.

VIGEANT: How does the ability for robots to think combine with the Internet of Things and with us as people? What will this mean for the future of jobs and automation?

ANGLE: First of all, the Internet of Things is a very broad term. It basically speaks to the fact that most devices in our world over time are going to be connected to information networks in the cloud, enabling them to share data with one another. Today, you see this in factories where machines that build things can now let their operators know how they're doing. They might convey, for example, whether they need service, how many units they've built or whether they need more material inputs. This connecting of machines within and among factories is making factories more efficient.

In the commercial world, buildings are getting smarter. When you leave a building, for instance, the energy required to heat the areas you used to occupy are no longer heated or heated differently. So commercially, using smart devices is helping us operate more efficiently.

In the consumer world, everything in our homes—from our stove, microwave and lights to our televisions, radios and robots—is all becoming connected. When you walk into a room, the lights can turn on. When you leave, they turn off, saving you money on your electricity. Our homes can take a little more responsibility in maintaining themselves, so that we can spend our time with more freedom of choice. If you want to go ahead and spend your day vacuuming because you enjoy it, fine. But that will become a choice in the very near future, not a necessity, because you can allow the house to manage cleaning the floors for you.

The promise of the Internet of Things is that more and more objects are becoming connected and automated to deliver a very sophisticated experience. One of the ramifications is that a lot more software will need to be written in the near future. We need more control systems. As that happens, the cost of connecting smart things to the cloud is going to go down, and the ramifications of low cost of connectivity is really quite remarkable. I recently read somewhere that diapers are becoming connected to the cloud, so your phone will go off when your baby wets their diaper, but before they start screaming.

From a career perspective, it means different skills will be required and in demand for a very long time, and we need to make sure we address that as educators.

Internet of Things
Internet of Things

VIGEANT: That connects with the idea of K-12 STEM (science, technology, engineering and mathematics) education. If devices will be connected in private, commercial and educational settings, and can begin to learn our patterns and preferences through artificial intelligence, what does that mean for how we teach science and engineering in schools? Should it be based on content, skills or both?

ANGLE: Those are all huge questions. Smart connected devices and integrating sensors with databases offers new ways of teaching. We need to educate people who can actually create, maintain, improve and enjoy these types of devices. Does that change some of the goals in education? I think so.

I'm not sure who coined the term post-knowledge society, but that is what we're living in today. I'm amazed by the younger generation's ability to look things up. The idea used to be that someone was smart if they could remember lots of things, but we need to challenge ourselves about the relevancy of that old definition, because these days facts are found easily. The more valuable commodity that's emerging is how we look for those facts. How do we synthesize those facts into solutions or into solving a problem with which we might be faced?

That makes critical thinking increasingly important. I'm not saying we should abandon fact-based education, because I think you need to know and understand enough about your world to give you the context to think, but we certainly need to emphasize that idea of bringing facts together. When we have a problem to solve, how do we come up with ideas that allow us to go and find facts, then combine them to form a solution? That's emerging as one the most valuable skill someone can have in the world today.

VIGEANT: Bloom's Taxonomy is an important concept in education. The idea is that there are different levels of thinking. Remembering, understanding and applying comprise the lower order thinking skills, while analyzing, evaluating and creating are the higher order thinking skills. Typically creating has been at the top, but we have recently reorganized this taxonomy to reflect the fact that analyzing, evaluating and creating are really equally important. The question is, should the goal of contemporary STEM education be to develop and use content with those higher order thinking skills with the goal of identifying and solving problems?

ANGLE: Yes. I think the new Bloom's Taxonomy and I are on the same path. I think it's interesting calling something higher order thinking, because it implies that some skills are better than others. This implies that you'll be more rewarded in society for evaluating than for simply remembering, and that's probably true. If I had my druthers and could wave a magic wand to control Earth, I would want to emphasize the skills toward the top of the chart. They are traditionally underemphasized and are increasingly critical to succeeding in today's environment, where the premium on remembering is relatively low. Although it's important to keep in mind that along with the ready access to facts comes the risk of misunderstanding or misuse of knowledge Having knowledge without understanding doesn't actually set you up well to evaluate, analyze or create.

But certainly the skills at the top are understated and the skills at the bottom overstated. I don't think STEM education is a synonym for this chart, but these skills are definitely important in a STEM curriculum. Somewhere along the way, science, technology and engineering have fallen off the traditional curriculum. Math less so, but even that is underserved considering the leaps and bounds by which its importance to the world is increasing today.

VIGEANT: So what you're saying is that while critical thinking is not unique to science and engineering, it is necessary to engage with them productively?

ANGLE: Yes. Critical thinking is central to STEM subjects but not unique to STEM subjects. I think problem solving is also important around social issues and challenges in our daily lives, but this type of thinking is certainly the bread and butter of STEM topics.

VIGEANT: One of the models we often see, and the model that reigned in education for decades, is the one in which the teacher is the content expert who models, explains and demonstrates, and the student is expected to show their proficiency by being able to recall, summarize and repeat. One of the criticisms of this traditional model is that there's not a lot of opportunity to exercise creative, evaluative and analytical thinking.

What kind of a paradigm shift you feel here? Is this a model we should continue with, or if not, what should we move toward in order to teach the types of skills necessary for innovation today?

ANGLE: I think that the traditional model does a reasonable job of teaching knowledge but not the thinking skills higher on your chart, because in this model the student doesn't get to engage with those skills in the same way.

When I'm interviewing someone for iRobot, the question is less "Where did you go to school?" or testing them on facts they've memorized, and more "What have you built that should convince us you're a good fit for iRobot?" We ask how they solve problems, where they go to solve them, what skills they employ to do so.

I think that in education, it puts a significant additional burden on teachers to shift from this traditional model to a model in which the kids are challenged with finding solutions to problems and given enough tools so that they actually can come up with solutions that are meaningful. It's a complicated mission, but I think that that's where I would like to see our education going. That's how we educate people once they're at iRobot. We give them stuff to do. We don't send them home with the textbook and say, "Here, memorize this textbook." We tell them, "Here's a problem. You have the background. You understand the basics of physics and engineering. Now figure out a solution that builds on those facts and skills." An innovative job poses questions and asks people to draw upon their personal experiences and upon accessible facts in order to meet those challenges.

So when we're interviewing, we look for people with the skills to solve problems in that fashion, which is what will make them good employees. Thus, if our educational system could teach those skills, I would certainly have a much easier time finding people who could work in innovative environment like iRobot. The traditional model does some things well, but it's very controlled to enable classrooms to have a high student-teacher ratio. A more open-ended problem solving model must be more hands-on, but that requires more resources, so it's harder to do. If a kid comes up with a solution to a problem, or part of a solution to a problem, that is actually quite innovative. So how do you grade it? It definitely creates all manner of challenges to our current educational system, yet it is the direction that I think we need to be going.

VIGEANT: What you just described is something we refer to as the next generation model of instruction. It's the idea that students are supposed to develop skills. In this setting, the teacher acts as a coach and helps students to develop those skills, then works to strengthen the connection between students and skills, as well as students and content itself. This creates an environment that offers the opportunity to develop and use the content instead of just trying to memorize it. Is that closer to what you think would be helpful in creating the kinds of employees you would hire at iRobot?

Next Gen Model

An example of a next generation model of STEM instruction

ANGLE: Yeah, I think this is absolutely what I'm describing. The challenge is that it asks the teacher to be a coach, allowing students to go far afield and simultaneously give them good advice to guide their progress. I think that this next generation model is absolutely the right way to go. But I also think it begs the question of how we get teachers the help and support they need to effectively play that role as coach.

One of the challenges to science is that it's a topic that is so profoundly broad that even if you devoted your entire life to it, you would only be scratching the surface of one small dimension. So this concept of "Hey, let's teach science!" is a deeply challenging prospect for a teacher, particularly if that's only one of the topics or subjects that they're asked to be proficient in. I think that's a case for bringing in some additional structure or support, perhaps in the form of someone else in the classroom who can work alongside the teacher. We need to figure out how to get there because I like your idea of teaching skills. I like that idea a lot.

VIGEANT: Thank you. So beyond science and engineering, we also have to focus on several other subjects as teachers: English language arts, math, the arts and so on. A good command of all of these is necessary if we hope to position our students to take advantage of opportunities—college, a terminal degree, career. With regard to hiring the right people for your company, which of these subjects and skills makes students trainable or adaptable in such a way that they're ready to come work for a company such as yours? What is really needed to succeed?

ANGLE: First off, I think the ability to solve problems and access the knowledge required to solve those problems is useful regardless of the career trajectory you're on. If you're a technician, you're going to be asked to service an increasingly wide variety of increasingly complicated pieces of equipment, whether it be air conditioning or heat pumps or whatever. There's no way that, even if you're brilliant, you're going to carry around all of the minutia and knowledge that's required in order to solve the problems posed in that career without accessing the information needed to create effective solutions.

Technicians need to be able to say to themselves, "Okay. The customer is complaining about this particular issue. I can go and look up what I'm supposed to do, gather that information and create translations from what the customer says and the observed behavior of the system to the solution." In other words, using the information they access that they do not intrinsically know, the technician can create viable solutions to problems that they may or may not have seen before.

These are the types of skills that you're talking about. Then you move on up the food chain to being a biochemist or a roboticist and it's the same challenge, just at different levels of complexity. There are different languages associated with the information that you're being exposed to. I think that if the goal of education is career readiness, then the approach of education needs to shift from an emphasis on rote memorization to an emphasis on using your language skills to allow the analysis of the problem at hand, the retrieval of the necessary information to solve that problem and the execution of that knowledge.

VIGEANT: In your own experience, where do you think you picked up those skills? What made you think "I should do this"? Was it in an elementary classroom? Was there somebody or something that really sparked your early interest?

ANGLE: Yes. I think that I was very fortunate. Perhaps it's the way that my brain works, but I've always been lousy at remembering. My spelling skills were always very poor. I was in remedial math, I suppose because I just couldn't be bothered to memorize my multiplication tables. Because of all that, this magical thing happened. In 5th grade, I was in my remedial math class and the teacher said, "You're bored," and put me in honors math. The teacher gave me a calculator and enabled me to get straight As in math from then on. I went from Ds to As, from remedial math to honors math, because I was put in a situation where my skills were valued. Rote memorization was somewhat important, but I was given a workaround.

I think that, to some extent, I was allowed to focus on problem solving skills rather than memorization skills, and this was the case throughout my educational path. Part of my problem solving skill set was learning how to convince a disturbingly large number of my teachers to allow us to have open book tests. This allowed me to avoid memorizing things, so I could look up the beginning of the formula, derive the rest of the formula on the test and answer the question. My career really benefited from these experiences, because it's based on developing very good problem solving skills in any given situation. I can figure out how to derive the answer from facts I can go and look up.

I think that it was this 5th grade teacher who understood that I could be smart, not by memorizing rote facts as well as the next kid on the block, but by deriving answers on tests to questions others could not solve. I learned there were other ways I could succeed, and that was valuable.

VIGEANT: It sounds like a hard work ethic was part of your success. You were determined to succeed, and so convinced teachers to allow you to have open book tests, etc. I'm reminded of a quote by someone named Angela Duckworth who's done a lot of research on an emerging area called grit. She says that talent is common. It is really heading in a specific direction with determination that predicts success. Is that an idea you'd agree with? Is that something that has been part of your experience in a way?

ANGLE: Well, keep in mind you're talking to someone who went six and a half years with this company before we ever had enough money in the bank to make payroll. So certainly, iRobot exists because of a particular determination to succeed. Grit is actually something we look for when we're interviewing people, and I'm familiar with the idea. Doing things of value, innovating and succeeding certainly involves a lot more failure than success. If you can get to that success, you can change the world in exciting ways.

I think that the stick-to-it-iveness is incredibly valuable in a student because, as I said, the problems that really matter—that are really challenging—are the problems that are difficult and make you work hard. And if you can successfully work through them, that makes you special. It makes you valuable, because you become qualified to work on the world's most interesting problems.

VIGEANT: Since you're a fan of grit, one of the other things that Angela Duckworth says is that rigor is where challenge exceeds skill. Everyone in the education world seems to think of rigor a little differently, though, so I'm curious about your take on it. Is rigor, in a sense, the root of innovation, or is it something else? Or is rigor just having direction and determination?

ANGLE: I think rigor is being in a situation where the challenges are not immediately obvious, and therefore require grit and determination to move forward. I also like the definition happiness as being challenged and succeeding. If you are increasingly able to encounter problems that to you seem very challenging, and yet you have enough of whatever it is inside you to solve it, that's exciting. It's also self-reinforcing to go and raise the bar when looking for the next challenge. May we all be so lucky as to have careers where we are challenged and yet can succeed.

VIGEANT: You have a product called the iRobot Create 2, a highly programmable robot that can teaches programming to beginners and advanced students. It's an amazing STEM resource, and you have others as well. So you're helping people in problem solving by giving them a resource that helps them develop skills. How can a tool such as the iRobot Create 2 fit into a classroom and a student's life?

ANGLE: This was spurred by thinking about what we, as a company, could do for the world beyond our products. We decide to try and promote and provide educational tools. We realized there as an obvious answer: We build 2.5 million robots a year. There are other educational robots out there, but not with the infrastructure we possess. If we could apply that mass manufacturing expertise to making an educational robot, we could probably make more robot for less money than anyone else in the world. That's what the Create 2 is. It's based on the Roomba platform, enabling us to take advantage of all of the economies of scale we already have. It is based on our current operations and can deliver, for a very low cost, a robust platform on which you can make really interesting durable robots.


iRobot Create 2

Making it especially beneficial for education is the fact that robots are some of the most exciting and inspiring things in the world today. They have a distinct ability to get kids excited about learning the skills that we're talking about. Kids are forced to ask and answer questions: How does this device work? How can I get it to do my homework for me? That type of problem leads to wonderful learning experiences, so the Create 2 is one of the tools in the iRobot arsenal that can help kids get more excited about STEM education.

Just how does it fit in the classroom? There are a number of different activities, programming languages and so forth, that have been built on top of this foundational platform. They tend to be targeted at late high school classes, after-school-type programs or college curricula. There are some things that you can do with the Create 2 at younger ages, but it is a pretty sophisticated platform on which to build some real robotic programs, which can make it hard to go younger. To address this, we are working on some robots that are easier to program. We also have programs to bring kids into iRobot, into our facility, so that they can talk to roboticists and to engineers, learn about what they do and get a first-hand experience of the creative and innovative processes that go into invention.

VIGEANT: This platform is something that students can access, engage with, program or reprogram, correct?

ANGLE: Absolutely. In fact, iRobot was built by a bunch of hackers that love invention and education, so that single robot we build—including every Roomba out there—has a serial port. If you want to, you can download the API [application programming interface, a collection of tools, protocols and subroutine definitions enabling the possessor to build software and applications]. This allows you to turn your robot into one that is reprogrammable. The Create 2 just makes it easier, and we pulled out the vacuuming system so that it will be less expensive. We also added some places to put additional sensors and computers as well as a nice top plate to which you can bolt things. These robots are ready to be reprogrammed, to be put to new and innovative uses.

We depend on teachers to provide and educate our workforce, so we want to give them tools to do so. iRobot is a 26-year old company, and we certainly intend to be around for a long time. I think that it's important to ask, "Are we educating students in such a way to enable students to get jobs?" From a particularly self-serving perspective, I want to do what I can to ensure that the next generation of roboticists is being brought to the schools and are going to be able to raise the bar from where we are today. I'm very happy to participate and I'm very supportive of what you're doing with KnowAtom.

VIGEANT: Thank you very much. In closing, what would you say to teachers who believe in what we've been talking about? I know in the past you've said robotics has had to overcome some skepticism as a field. So if these teachers appreciate the big-picture value of STEM education, but find themselves surrounded by skeptics who see it as just another fad or acronym, how might they proceed? What advice would you give them?

ANGLE: STEM is an acronym and it stands for science, technology, engineering and math. I've said this once before. When did it become okay to not teach science in school? The amount of science that our kids are getting in elementary and middle school, and even high school, is beneath what teachers and almost all students expect. I'm blown away as I look at middle school curriculums and there is no science. What? That's not right. Even in Massachusetts, one of the best places in the country for education, this is the case in many, many public schools.

I think science has become politicized, so now when schools are graded, the scores on science no longer count based on the national standards. Therefore, if a school wants to do well according to those national standards, they no longer need to focus on science. So then it doesn't get taught, or gets de-emphasized, and there's a huge downward spiral.

The current focus on STEM is really just fighting back against the momentum in our educational system that is de-emphasizing science. STEM saying about that this is not okay. The future and competitiveness of our nation is going to be determined by the skills of the students we're producing, and the most important of these skills are going to be the STEM-oriented ones. Maybe I'm reacting too strongly to the question, but I just don't understand when it became okay to stop teaching science. 

So I would say that if you're surrounded by skeptics, the real question is, how was it ever okay not to teach STEM subjects? Creating students who are ready to take on careers and help our country continue to be a leader is dependent on us fixing this. I would say stand strong.

“Growing up, I wanted to be an inventor, solving problems that would help people have better lives. Every day at KnowAtom is an opportunity to invent solutions that give thousands of students and teachers a better experience doing science, engineering, technology, and math (STEM). Providing educators with professional satisfaction and students with the opportunity to understand the world we live in is my way of helping people have better lives.”