As part of our ongoing AI In your Community series, I sat down with Dr. Maja Matarić, a computer scientist and roboticist known for her work with human-robot interactions, and specifically for her pioneering work with socially assistive robots. During our conversation we discussed how to identify real-world problems that people actually need to have solved, finding your passion, and how to convince others of the value of your work.

Dr. Matarić and Chili, a robot used for multiple projects, including helping 1st grade students learn about making healthy food choices.

Tara Chklovski: What inspires you?

Maja Matarić: I don’t think I always knew what inspired me, but have now finally realized that what really motivates and drives me is helping the underdog. I realized I am just a person who really wants to feel like they’re helping someone – and that’s how I came to discover assistive technologies, because their whole point is about helping people, and for me it’s really about giving assistance to someone who really needs it. I want to reduce someone’s pain, ideally quite literally and if not literally, than figuratively. And so, that’s what motivates me.  It  could be inventing ways in which robots can help people with special needs, or helping my graduate students find their own path, or helping girls realize that they’re just as smart and just as good at math and just as capable as anyone else. All of these things are incredibly and equally compelling. There’s no shortage of compelling problems.

TC: Solving the right problem is kind of hard because problem identification is hard. What advice would you give children and parents as they try to find a problem in their community to solve using technology?

MM: Ah, that’s a great question. The thing that I always say to all of my students – a day outside of the lab  is worth a year in the lab. For instance, I know some students are very interested in helping the elderly, so I tell them to go to a retirement home or a nursing home, where they will immediately witness so many challenges that people are dealing with every day. That’s just one context, but it could be anything.  Talk to people who are in the context you are interested in, and they will immediately share with you “I wish I could do this, I wish I had that”.  Getting out of one’s own context and getting immersed into someone else’s is really the only way to find out what the real problems are.  It’s that old adage about walking a mile in someone else’s shoes.  You can’t do that by just being an observer, you have to actually do it yourself.

People who invent and develop things often do it out of curiosity or just because they can and then are surprised when their invention/product  isn’t something that other people need or want. And I don’t know why it’s surprising at all because people who are doing the inventing and developing this are not typically representative of the people who will be using whatever is being developed. Don’t think that you are really representative of anything other than you. Go out there and talk to someone very different from yourself. It could be your grandma or your aunt, it does not have to be a total stranger. But go talk to someone else. Don’t just come up with ideas in your head because unless your target is people exactly like you, you are not going to get it right.  When you get people with real challenges to tell you their needs and wishes, you will learn a lot and find endless problems worth working on.

TC: What do you think defines a good problem?

MM: Good problems are not faddish, they are timeless.  This means they are hard, but also that they are meaningful, and are not going to become irrelevant like fads do. I think problems people encounter in the real world have great meaning in them.  So, that’s why I always advise students and startup founders alike to get out there and talk to real people who have challenges to find beautiful worthwhile problems to work on.  I am not very impressed with people who develop things just because they can or because they can make money on them.  I admire those who care more deeply about helping others through meaningful work, which may be harder, but it’s more important.

TC: Going a little bit into the process of problem solving – what do you find difficult and how do you overcome that difficulty?

MM: Solving difficult problems takes a team, and it takes thinking outside of the box. If it didn’t, someone would have already solved them. Getting people to listen with open minds and take risks is much harder than anyone assumes. It turns out that every time you want to do something in life you really have to sell your idea to someone, often a lot of people. You have to do a pitch, and you have to get people to buy into it. Whether you’re trying to get funds from an angel for a non-profit or a venture capitalist for a startup, or a federal grant for research, or your boss for a new project, or support from your family for a new endeavor, you need to be able to convey to enough people that what you want to do and what you need to do it are meaningful and worthwhile and doable. That’s a skill that we all need to keep honing life-long.  Unfortunately there are numerous great ideas that never got support to be developed, and that is sad. So learn to pitch your ideas: practice, listen to feedback you get, and keep trying.

TC: So that’s great, great, great advice. And actually, a part of this AI family challenge is that the families are going to be showcasing their product and their designs to the community. So, what key things have you learned in trying to convince other people of what you’re working on?

MM: I think there are different approaches to convincing people because there are profoundly different people, but in the end you have to be true to yourself.  I tend to be very passionate. So I end up being very honest about what I believe in, and pretty pushy about it, because I really can’t help myself.  Some people resonate with that, and they provide key support, in various forms: emotional, social, economic.  Other people have more subtle styles, and some figure out what their audience would like to hear and then craft their message that way. That approach can be incredibly effective, but just isn’t my way.   But whatever your way is,  the thing about getting support  is that you keep pushing, keep pushing, keep pushing, keep pushing…it’s a lot of pushing. That’s my big message. It’s always a lot of hard work. You never get what you deserve just because you deserve it.   Meaningful things take hard work.

TC: How do you think AI specifically will help strengthen societies and communities?

MM: Well, it’s very tricky with AI in particular because AI is a very general technology, so it could be used in various ways, some good and some not. I think there are going to be wonderful deployments of AI that bring together different types of data in order to enable complex modeling, prediction, and especially personalization of things and services for people, which is really wonderful, because it can serve to support a diversity of needs. Hopefully healthcare and education will get better,  as both are in desperate need of significant improvement; personalizing content and services for specific users/learners/patients has the potential to significantly improve quality of care and education, leading to societal-level outcomes.  So there is great potential for meaningful and impactful uses of AI, but there are also shallower uses (we are already seeing many of those), and worse yet, uses that invade privacy, and lead to pressure for conformity rather than personalization.  Any company or agency that has a great deal of data and can develop complex models, can in turn use those  data to label, classify, and target people.    I am currently much more concerned about those uses of AI than I am about AI replacing human jobs, though that, too, is a legitimate reason to worry, and a great reason for numerous young people to go into STEM fields so they can shape the future of technology and its uses.

TC: So, what do you think is the best way for children to learn more about AI?

MM: Start coding as early as possible.  I am happy to hear that now there are coding camps for preschoolers.  Unfortunately, at the same time, most high schools still have overly simplistic computing classes, because K-12 curricula have not caught up or kept pace with the need for STEM education. So I expect that long before school curricula catch up (because that takes state and national policies and funds), informal after-school and summer programs will be the best place to get some skills, along with great online resources, such as

Coding is everywhere. You can control devices all around you, including robots, by creating action in the world by knowing how to program. It’s like learning the alphabet and it should be taught just as early and just as dedicatedly.  Kids are amazing and can learn coding quickly, and once they start thinking computationally, they can start to see the world in a different way, a way that lets them create systems and technologies.  That’s wonderful because once they “get it” then they can “drive it.” They need to take control of their future.

TC: Exactly. And that’s what we’re trying to say, they are getting exposed to all these cool technologies, but there really is a powerful piece that you have a role to play in here.