Camille Morhardt 00:37
I’m Camille Morhardt. Welcome to today’s podcast, What That Means: AI Literacy in Today’s World. I have with me, Tara Chklovski, who is CEO and founder of Technovation. Welcome to the show, Tara.
Tara Chklovski 00:47
Thank you, Camille. I’m excited about this conversation.
Camille Morhardt 00:52
I’m really glad to have you on. You’ve been written up. You’re well regarded and well known in the land of AI education, and including you were recently written up by Forbes for your very ambitious goals of educating or reaching 25 million young women in the next 15 years. Before we get started, maybe you could just let us know what is Technovation? I know it originally was founded with a different name, and then we’ll dive into the conversation.
Tara Chklovski 1:20
Yeah, so I started Technovation, which was called Iridescent, almost 18 years ago. The goal was to bring the world’s best technologies and the best education resources to the world’s most needy communities, to the most vulnerable populations. And over time, it has focused in on girls and women, because there’s no country in the world that has gender equality, and they’re the ones who are on the forefront of many, many, many hardships. And so that’s what Technovation is.
Over the years, we have narrowed into one particular form of programming, which has shown to have long-term impact, which is a 12-week competition where girls work with mentors from industry. They find a problem in their community, and they develop either a mobile app or, more recently, an AI prototype, and actually develop a business plan to take it to market. This three-month experience has shown to have life-changing impact on these girls. Based on that data, we decided we need to bring this to many, many more girls and young women around the world. We are partnering with UNICEF to be able to do that, and our goal is to reach 25 million young women.
At the moment, there are only three million women in technology who are professionals, and we hope to double that to six million by 2030 or so. The number of women who are driving innovation in the tech sector dramatically needs to increase, and I think that’s a very, very powerful risk mitigation strategy, especially when it comes to AI.
Camille Morhardt 2:56
And we were talking about AI literacy, just generally speaking, and maybe even it’s more broad than that–maybe it’s like technology literacy in today’s world. It used to maybe be familiarity with a computer, or then familiarity with the internet or access to it. Do you have a sense of a definition for Technovation, and then also just, is there a generally agreed upon definition for technology literacy or AI literacy in today’s world?
Tara Chklovski 3:27
Yeah, and I think the education world has been grappling with that, because it’s beyond just the memorization of terms. So I think computational thinking is one interesting lens. For us, it’s really about going beyond understanding how it works and being able to apply it to complex real world problems, because that’s when you really begin to grapple with the technologies. That’s when you really understand its limitations, and that’s when you really are able to bring your own lived experience and come up with something new, and not just learn how to code because it’s an important skill or anything like that, but it’s more about build a mobile app that is solving a real world problem, so it’s a tool.
So in 2010, it was mobile apps. We started doing AI education in 2016, and we just finished our World Summit Care. I think we had about five or six, out of the 15 finalist teams, who used AI in their solutions. I was just blown away, because they are learning how to use large data sets. They’re learning how to have varied representation in their data, and then they’re training machine learning models and using all sorts of different algorithms to be able to come up with better higher prediction and higher accuracy.
I would never have imagined that, in 2016, these girls would be able to do this. Girls as young as 11 years-old, they’re using Kaggle data sets on lung cancer in India so that they can provide better guidance to patients who may be at risk, because so many of these data sets exist now as open dataset. And so data science is underlying a lot of this, and so it’s probably one of the components of the AI literacy, but to me, the very exciting part is, what are these girls solving with these kinds of building blocks?
Camille Morhardt 5:21
Yeah. Could you give us a couple more examples?
Tara Chklovski 5:24
Yeah, so one example was to help teenagers assess mental health issues, and so they ran a survey first, and they got very few respondents. So then they went and actually found an open data set of a lot of tech snippets, and they did a sentiment analysis on that, and then they trained their model on if a child were to sort of say how they’re feeling, t hen, it would kind of give them guidance and feedback, and so that was one example of sentiment analysis and figuring out, “Okay. Our current dataset is not large enough. Do we have that kind of accuracy?” So they went and found one.
Another one from a team in Canada looks at chest X-rays to look for this symptom, your risk of cardiovascular disease, and so it’s called cardiomegaly. I didn’t know what that term was. Then, they trained their model on these chest X-rays and came up with the way to sort of predict whenever you’re at risk.
Another one from India created an earthquake prediction app, because in India there isn’t an app, or in many countries around the world. I mean, this weekend we had this massive earthquake in Afghanistan, and so there’s no prediction system. And so she’s using USGS data, and then overlay it, and then use random forest algorithms to sort of come up with a decision tree to say, “What is the likelihood of something to happen, and then what is a safe exit route?” So it connects onto maps, and helps you sort of figure out, “Okay. What should I do when this is happening at this moment?”
So just very, very, very interesting use cases; and that’s the power of having a diverse group of people learning these tools and applying it to their problems. So we don’t stop at, “Oh. Learn about how this machine model works. Learn about how ChatGPT works,” but it’s like, “Use it to solve a problem that you’re facing.”
Camille Morhardt 7:16
Can you give us a sense of where all of these girls are from and how big are the teams? I’m also interested in, I think you pair them with mentors, and so what would a typical mentor? Obviously, this is going to require some kind of technical expertise from a mentor as well to help guide on size of data sets and whatnot.
Tara Chklovski 7:35
Almost 50% of our mentors don’t know how to code, and they’re not technical, and they participate in this program because they are learning, because we have a free curriculum that walks you through from the very basics of, “What is AI, and is my robo rock or my robot vacuum cleaner an AI or not?” all the way to actually developing a functional prototype and understanding what are good data sets and not.
So the girls come from all over the world. I think just because World Summit is fresh in my mind, one of the teams was from a very remote region in northern Kenya, which is a fishing area, and they created an app. It didn’t have any AI in it, but basically providing tide prediction information to fishermen. Because of climate change, lots of things are changing, and these kinds of basic information is not available to a lot of people in the world. So, we work in roughly 120 countries around the world, and many times I have to look up these countries on the map. One of my favorite alumna who actually has launched an actual business that’s running, is from a Caribbean Island called Dominica. It has a very small number of people. I don’t want to get the number wrong, but it doesn’t have a university, so she’s going to a virtual university. So just the whole range. Our big countries are, of course, US, Canada, India, Nigeria, Mexico, Kenya, Brazil, Chile, Spain, but a lot of countries, and I’m sure we can look up and see if we have girls there.
Camille Morhardt 9:03
So I want to switch. I want to come back to the projects and the people, and some of the concerns too, are we thinking about ethics and privacy? But before we get back into those, I want to know a little bit more about you, because you are all but dissertation for a PhD in aerospace engineering, and specifically on, I think, the aerodynamics of bird flight, which is fascinating, I think. And that was from University of Southern California. So tell us how did you decide to pivot and do this instead?
Tara Chklovski 9:41
Well, that was such a long time ago, and good job on the research. I think I’d always wanted to be either a pilot or build airplanes, and especially those inspired by birds, because birds are incredibly beautiful and so efficient. I think when I was eight years-old, I grew up in a pretty poor area in India and community and family, but I had an old copy of the Popular Mechanics, and it had Paul McCready on the cover, and Paul McCready was the father of human power flight, and he made the first solar-powered airplane. He made the first solar-powered car, race car. And so I was like, “I want to work in this company,” and so I got a degree in Physics, and then I got a Master’s in Aerospace in Boston University. Then, I slowly made my way to southern California where his company, AeroVironment, was.
I was in a PhD program under an advisor who, at that time, was the only person in the United States studying bird flight. I think I didn’t really know how to choose an advisor. I didn’t know that he had never graduated a student successfully, so I really wanted to study seabirds because they are incredibly efficient, they can go for very long distances. I got an internship at AeroVironment, and I was working there. I was the only woman on the research team, and at that time they started to slowly switch towards drones. They were the first company to start making drones that went into the Iraq war, and that was so far from what I had imagined, and I didn’t want to work in a big aerospace company. There wasn’t a aviation startup industry as there is now, so it was a real time for me to reset.
I spent, I don’t know, maybe I’ve forgotten how many years I was in the PhD program, maybe four or five, and there was no other option. I didn’t feel like I wanted to just quickly finish and go get another job, so it really was going back to the drawing board of, “What can I do in the world, and what are some big problems that resonate with me?” My family, the women have a history of, when there’s a major change in their lives, they start a school, because in India, almost every country, education is an acceptable profession for women, but my grandmother started a school when she was 60 years-old. She cut down a tree and made it into one table, one chair, and opened up a school. The school is 45 years old now, and thousands and thousands of children have gone through it. And so my mother, she was a doctor in the army. She left the army, and then she was running a school.
And so education was definitely a key part in my mind, but as one of the most powerful solutions we know to address inequality and injustice. I just always feel that where you are born is no credits to you, and where you’re born determines so much of your life. We did nothing to earn it or not earn it, and I think that should not determine a person’s faith anymore, and education can change that. Internet can change that, and a combination of that can change that, and those were sort of the underpinnings of how I started Iridescent.
Camille Morhardt 13:03
Thank you for sharing. That’s really interesting. Do you feel like there’s an element of technology now? And maybe you’re going to say AI or maybe you’re going to say large language models. I’m not sure, because it allows coding to occur just in language as opposed to having to, I shouldn’t say new, because a lot of people would have to learn English in order to use some of the LLMs. But what is your take on whether there’s a new or an emerging concept in technology that might really help make it matter less where you were born?
Tara Chklovski 13:36
I think education. I think the people-focused strategies are really the key there. I think I’m not a very big fan of talking about the digital divide, and the reason is that I think it’s a very solvable problem, and it is one of the solvable problems–or it is one of the things that you’re making the most progress on. So on the UN 2030 agenda, there are 36 indicators, and there are only three indicators that we are on track as a world. Those three indicators are access to electricity, access to internet, and access to mobile phones, nothing else. And so people will get access to internet, people will get access to mobile phones, because guess what? It’s in the best interests of corporations to do so, and for the consumer, it provides immediate value, so it’s not a very hard problem.
I think the harder problem is things which are people focused and are not as glamorous, sexy, or the silver bullet part, where you say, “Okay. These number of people need to get access to phones,” and then somebody cuts a check and you get the phone; so the harder problem is, “Well, you need to train all these teachers who then need to go and have these kinds of more interesting ways of teaching these children, and then you have to assess the impact, follow up, and make sure there’s no corruption.” So I think focusing on bringing these powerful technologies to girls is a very, very, very important strategy for addressing inequality today.
I think one of the problems is that, too often, girls’ education is limited to primary and secondary education, and people think in sequential. “Why are you teaching them about AI when they’ve not finished like sixth grade or seventh grade?” And I love what Seymour Papert used to say. There would be a world where a three-year-old could interact with a computer, and ask and say, “Show me a picture of a bear,” without even knowing how to read or write, and that totally happens now. So there’s no reason to have a sequential way of saying, “Oh, girls need to go through primary and secondary before learning about computer science.” And too often, many girls focused initiatives and organizations are run by women themselves who are tech phobic, because they went through a lot of the same systems, so they’re not as interested or excited about bringing AI into this. AI feels very scary, so I think, then, you have sort of a vicious cycle where you’re not bringing in these kinds of new technologies and new ways of learning and thinking into these very vulnerable populations.
Camille Morhardt 16:12
Let’s talk about bringing AI then into education, and what are some of the reasons people are afraid of it? Because surely some of those are valid, right? I mean, how are you addressing teaching, helping people understand, or interact with technology that has big implications for privacy and ethics? How are you bringing those components in, or are you purposefully not bringing them in and saying, “This is a technology, and that kind of thing has to be handled elsewhere”? How are you dealing with that?
Tara Chklovski 16:44
I think just education, like being very open about it, and that’s why we have a very detailed, three-month curriculum that teaches the adults, really, honestly. The girls are not worried about these things, but we teach them exactly about how a large language model works. How do you train a model? What does a good data set look like? What’s wrong with this? So I think education is the only way to combat fear.
And I think I’ve heard too much of the same kind of argument, “Here are all the risks.” Well, yeah. I mean, every kind of technology has risks. The only way to combat that is to change the makeup and format of the teams that are behind this to make sure you have a very, very, very broad set of people who understand how this works. So they’re asking the right questions, and they’re using these technologies in the right way, so it’s not going to stop and it’s not going to slow down, so you better catch up.
Camille Morhardt 17:38
What about, you mentioned that a lot of the girls are looking at open data sets so that they can build models with enough data to make a difference, so what is your perspective on that? I’ve kind of heard multiple perspectives with different people I talked to about how much data sets, there’s a privacy concern. Obviously, anything released would be, in theory, anonymized or some kind of privacy?
Tara Chklovski 18:04
So to be completely sort of clear, the girls, they go through this program once mostly, and when they create a prototype, they’re still in school. Even when they win thousands of dollars for the winning app, they’re still in school. So we give the funding not to launch their business, but for them to continue their learning and deepen, and then many of them come back year after year. As you come back, you get a deeper understanding of the problem or a deeper understanding of the tools, and that’s when you begin to like, “Oh my goodness. I never thought of this. I never thought of that.” So we have quite a robust set of curriculum, but they’re not fully processing every part of it. That’s why you study English for so many years, or grammar, literature, or whatever, and you keep peeling the layers of the onion.
In terms of open data, I think all of us are also grappling with that. What does it mean to own something when maybe a lot of people have contributed to it, and knowingly or not knowingly, but again, as an educator, I think our job is to present the problem and to help girls think about it. One of the things that comes up very, very frequently is, initially the girls would want to kind of patent their idea. I was like, “Guess what? You are here because people have donated their money. Corporations have funded this program. Thousands of volunteers are volunteering their hours. This program wouldn’t exist without generosity and kindness, and the right thing to do is to take your intellectual property and bring it back into the world and make it open so others can build on it.” That’s the viewpoint and philosophy that I hope the girls are walking away with, where rather than, “I’m going to keep my idea,” because there are plenty of great ideas. The hard work is in the execution, and maybe the main thing is don’t hurt anybody, and we have a pretty detailed ethics curriculum, but some of these questions are very, very tough. I know a lot of these companies haven’t figured that out either, so the right thing to do is to discuss and to learn.
Camille Morhardt 20:09
What are some of the most interesting questions that you get, or that either the girls ask, that the companies ask, or that NGOs are asking in this space? What are people still sort of perplexed about or wondering how to approach in technical education, AI education?
Tara Chklovski 20:29
I would say people are responding in kind of a fearful mode, where let’s just create a very detailed curriculum and give it to students. I don’t think that’s the right approach. So UNESCO did a pretty deep report on the state of AI education last year, and there are about 16 countries in the world that have adopted AI education as part of their national curriculum. And all of it is just a bunch of facts, or if anything, it links to Andrew Ng’s course on machine learning. Now, which kid would follow that course and get anything out of it, right? I think the key to think about is not from a place of defense or of fear, but more like, how can we empower students and young people to use these very, very powerful technologies to tackle these enormous problems that we are handing to them?
We are giving them climate change and severe inequality and not that many solutions, and I think AI is one of the positives. I’m not a complete rosy picture painter here, but I just think that we underestimate children. I see this all the time, and girls tell me this all the time that, “My parents never thought I was capable of this,” because they never thought that their daughter could get up on stage and talk about the AI model that she built, and she’s only eight. So I think that, to me, shows that we are severely underestimating young children, and we shouldn’t just stop at teaching them about how it works, but actually challenge them to say, “Okay. Use these tools.”
Camille Morhardt 22:04
Do you think that large language models specifically are changing the game as opposed to other kinds of AI programming that could occur?
Tara Chklovski 22:13
It’s interesting, yeah. So we were one of the first organizations to use chat GPT in the first week of December last year, because as soon as it came out, I was like, “Oh my goodness. This is going to have so many implications.” And so we rapidly created a whole series of resources and pushed it out into the community, and so we got back quite a bit of data because we were able to act so quickly. What we found was that most of the teams, of course, were using it for ideation and coming up with new solutions, but the second most common usage was translation, which was very interesting and not something that we had kind of expected. Our curriculums actually used block-based coding platforms. What was surprising was so many teams used that to ask for coding advice. At that point, I was like, “How are they using this for block-based?” But I guess it was teaching them what should be the algorithm.
So I think we see this all the time, right? Users surprise the inventors of the platform, but I think that coding will become easier and easier, but the need for computational thinking and the need for understanding how these models work are critical, because the tool itself is only a tool, and that tool still needs to be deployed into the real world. Only when you understand how it works, do you understand the negative implications or the positive implications–and especially in the communities that we work with. So I think that, even though coding could be done for you, I think there’s a lot of value to teaching children about how these things work and having them do it themselves, so that they could come up with better ways.
Camille Morhardt 23:51
And right now, actually, your organization is accepting new applicants, or can you just describe a little bit if people are interested?
Tara Chklovski 23:58
Yeah, absolutely. We welcome anyone who has a daughter, a girl who’s teaching students to join. You can go to technovationchallenge.org and sign up to be a mentor, to be a coach, to be a judge. We need lots and lots of volunteers. If you want to be a club ambassador, all of these take different amounts of time. Then, of course, for girls, it is the organization and the program that has research to show that it dramatically changes a girl’s sense of self-confidence, her problem solving skills, her abilities to work in a team, her ability to innovate all of the skills that the future needs.
Camille Morhardt 24:35
And when the girls get together in teams, do you sort of help form the team? Like a girl gives a submission, gives her age, gives her interest? Are they like teams from around the world or pocketed together?
Tara Chklovski 24:45
Absolutely. So I think the app that I was mentioning about the mental health one, some of the girls are in the US, and then one team member is in the Philippines. So yeah, so you can find, based on your geographic location, who’s looking for a team, and then you can also connect with mentors similarly, where you can see by expertise or time zone. And Technovation is an unforgettable experience, so I highly recommend everyone to try it.
Camille Morhardt 25:12
Thank you very much, Tara Chklovski, CEO and founder of Technovation. I really appreciate your time today.
Tara Chklovski 25:18
Thank you, Camille. Those were good questions.