Jennie Brooks, host of the Unstoppable Together podcast chats with Catherine Ordun, Booz Allen executive advisor and AI engineer completing her PhD in generative artificial intelligence. Tune in as they discuss how to make access to the technology itself and education about AI more equitable, as well as how diversifying the generative AI workforce leads to more equitable outcomes. For more information on the programs and resources Catherine mentions in the episode, use these links— The AI Education Project: https://www.aiedu.org/ Deeplearning.ai: https://www.deeplearning.ai/ Mark Cuban Foundation's AI bootcamp: markcubanai.org
Jennie Brooks:
Welcome to Booz Allen Hamilton's Unstoppable Together Podcast, a series of stories that unite us and empower each of us to change the world. I'm Jennie Brooks with Booz Allen Hamilton, and I'm passionate about diversity, equity, and inclusion. Please join me in conversation with the diverse group of thought leaders to explore what makes them and all of us unstoppable.
Hello, everyone. Welcome to the Unstoppable Together Podcast. I'm Jennie Brooks, and today I'm excited to be joined by Catherine Ordun. Catherine is a Booz Allen Executive Advisor, an artificial intelligence engineer completing her PhD in generative artificial intelligence. Catherine, welcome to the podcast.
Catherine Ordun:
Thank you. Glad to be here. Excited for the conversation.
Jennie Brooks:
Same. So everywhere you look today, we're seeing artificial intelligence in the news, in conversation, in our workplace, in our children's school. Tell us a little bit about the generative artificial intelligence, and how does that compare to other types of artificial intelligence?
Catherine Ordun:
I think in the past seven months, we've seen just a huge revolution when it comes to the general public getting a real understanding and gist of AI. And I think we owe a lot to generative technologies like ChatGPT, for example. And we've seen with ChatGPT some pretty stunning types of outputs that make it really, really different from some of the predictive analytics that we're used to, and there's a couple of reasons why. First of all, these generative AIs, they work in a completely different way than a lot of what we would call supervised machine learning that just output predictions for loan prices or classifications of if this is a dog or a cat. And these generative algorithms, you can think of a ZIP file, so when you're zipping up files together you're compressing the file and then when you want to open the file, you unzip it.
And so this is a really great metaphor for some of these generative algorithms because what's happening is that they are encoding really complicated patterns from images, video, text, and then they're decoding it. And along that process, they're learning about how different features or different patterns combine together in really unique ways that then allow for us to be able to generate completely new outputs. So this is not so much prediction as it is new combinations of patterns that, to our ear or to our eye, make human sense. And this type of technology has been around for about a decade, but I think it's because we've got really, really fast processing computers now that make this really possible, and then obviously we have a lot of data, like ChatGPT is trained on the entire internet. So it's a completely different revolution that we're seeing, but, Jennie, a lot of it is because these algorithms are very different.
Jennie Brooks:
Tell me a little bit about your choice to earn your PhD. What's your background and what does that educational track look like?
Catherine Ordun:
I have a unconventional background. A lot of times when people say, "Oh, you're doing your PhD, it's in applied computer science, you must have a computer science Bachelors." And that's not the case actually. I was talking to someone the other day, and we are chatting about how my undergrad is from Georgia Tech, in applied biology actually. And from biology, I went into public health at Emory, and then from there I got a scholarship to get my MBA at George Washington. And then I'd always wanted to do a PhD, mainly because I was really interested in research. And then quite honestly, Jennie, it came at the right time, the right place, the right professor, and also the right time of my personal life to make that big leap. So I followed a little bit of a unconventional path, but for someone like me who's a working mom, and then also just a person who was juggling a lot of personal demands, it was more about, is it the right time for me?
Jennie Brooks:
I love that because a lot of the conversation we have in this space is around the "unconventional paths", and encouraging folks to adapt, and take on new skills, and come from the growth mindset, so that all lines up nicely. And I'm thinking back to a previous podcast discussion with John Larson, who's also one of our AI leaders of our business, and we were talking about education and AI, and referencing your "alternative path". One of the things we know is that, with the pace of technology today... How do you think about lowering the barriers for AI literacy, and how do we open up that aperture to make sure that education, AI education is more equitable?
Catherine Ordun:
This is something that has been, in my opinion, very much of a gate-keeping exercise for a very long time, for multiple reasons. I think this generative AI revolution has brought AI to the masses. I mean, we saw with ChatGPT, it's the largest user adoption of any app actually in history. And so this has really allowed artists, content creators, students, to get a sense of, "Oh, what can AI do for me?" So in the past, it's been where you had to go to Coursera. You'd buy a course, you'd go through it, it's pretty dense. You might be a computer science major, then you'd go to grad school. So there was a lot of pathways that were very structured and academically focused.
And now what I'm seeing is, for example, with teenagers, I've been pleased to be on the board of advisors with a Mark Cuban AI Bootcamp, the Mark Cuban Foundation for almost two and a half years now. And that group is phenomenal because it exemplifies what I think we're going to see a lot in AI, especially when we are opening the aperture for younger folks. So with the Mark Cuban Bootcamp of which Booz Allen, we're going to host a camp in October, it's really geared for low-income high school students who actually may not even have their own computer. And the goal here is to really open up to the underprivileged, underrepresented high school groups. And some of the students will start as young as the age of 14, and they'll be accepted to the bootcamp. It's completely free, they even provide the computers for them, and working with them, I've helped to review, with a whole group of very brilliant people, their content.
And the goal is really, how can we get teenagers who are about to go to school, about to go to university, excited about AI? Let's not overload them with math, and equations, and code. Let's make it accessible, exciting, to really spark the imagination about, "What can AI do for me and my community? How can I use it to better my world and to better my experience of life?" And I think that's really exciting. And this is something that's been going on maybe for the past three or four years, but it's really exponentially taken new heights.
Jennie Brooks:
That's really exciting. Thanks for sharing. We've been talking about skills gap, but we also know that there's a gender gap in this field. According to the World Economic Forum, only 22% of artificial intelligence employees globally are female. You referenced a little bit about your path, what should we consider when we think about the gender gap in terms of careers in AI?
Catherine Ordun:
That number is even more dire when you get into the sphere of hardcore AI research. It's only 12% of the scientists who are publishing work in AI, only 12% are women. And that's startling because we now see an amplification of bias in AI all around us. Obviously, a lot of controversy from facial recognition where facial recog models are better at predicting white men, and very, very bad at underrepresented minority females. And then we see it even now actually with generative AI. I do a lot of work when it comes to generative imagery. It's not really a fault of the algorithm, but it's collecting and sweeping information from the internet. And the internet, as we know, is not the most pristine, clean place for unbiased and unfiltered thoughts. So we see that bias amplified in the imagery.
For example, if we write a prompt to give us AI art about a woman, it's usually Caucasian, and she usually looks like a supermodel. So this is not very representative of our population. And so I would say that to really have folks understand that they can enter this field without any gates is to follow something unconventional. There is not a one-size-fits-all when it comes to AI. As a matter of fact, to get into AI, you don't have to be a programmer, you don't have to be a software developer. There are so many things in AI that we need people for, from content creators, to program managers, to diversity advocates. I mean, AI is eventually going to infiltrate many, many aspects of our life, and there are so many economic and social ramifications. If we don't de-bias the field and level it out by introducing more folks, then there's a potential for it to go awry.
Jennie Brooks:
So how do we de-bias the field? Why is it that more women are not coming through at the outset? Because you're coming from biology-
Catherine Ordun:
That's right.
Jennie Brooks:
... what can we do collectively to pull more women in?
Catherine Ordun:
So I think it goes back to that gate-keeping posture, that's very, very healthy and very toxic. So I mean, there's some that are societal stereotypes that keep being amplified in our faces. We see TV shows like Silicon Valley, it's a bunch of tech bros, it's very humorous and funny, but you'd be surprised at how much that still really registers with society. And that there's a lot of content creators out there on TikTok, on YouTube, blogs that you read, and they're majority men. So this gives off this feeling that AI equals math equals computer science. Maybe me as a woman who's a communications major, maybe that's not really for me. But the jewel of it, the secret is that there is a misnomer to believe that AI is only about math and computer science. Really, to de-bias it, women have to have the confidence and the mentorship, the community, to be able to come at AI at many different angles.
Our world is not just math and science. We have communications, we have content, we have art, we have economics. These are all areas that we need people for in AI, and to really carve that unconventional path and find courses that you think are necessary for your growth, for your area of AI, I think is a part of the secret recipe. Don't follow what everyone else is doing.
Jennie Brooks:
What can we as mentors do to help people pull into this field? Or if people are wanting to make a change and are concerned about some of that bias, what are some of the resources out there, or some of the strategies that we can play to bring people to the table?
Catherine Ordun:
I think, again, us as mentors, us as leaders, we have to de-bias our own thinking first. I think that we come preloaded with some stereotypes about "what it takes" to get into AI or machine learning, and I'm afraid that some of the stereotypes are just amplified by media. So first of all, I think we got to bias ourselves, and we have to open the aperture about what we personally think is possible with AI. So to that extent, I get a lot of questions from high schoolers because I have an account on TikTok and I talk about my research, college students, obviously in my lab with my PhD work, I get lots of questions from people who are asking, "How can I get into AI?" And I could always give them the rote standard run-of-the-mill, "Look at this book on Python. Go to this Coursera website." It's like you're inventing your own curriculum, and that becomes very cumbersome because again, that one size does not fit all.
Now, that might be right. If you want to do AI math and AI computer science, you're going to have to maybe follow that trajectory. But AI is not an exclusive field to only math and computer science. So it requires, I think, all the leadership and the mentors to reach inside and to ask ourselves, "What do we know about AI? And now this person's coming to me, she wants to get into the field, what really suits her? What is her passion for AI? What is that niche, and how can I give some advice to direct her to those resources that she's passionate about?" And it may not be math and coding.
Jennie Brooks:
Awesome. Can you share with us just a few of the resources that might be helpful for folks, whether they're a mentor, or whether they're wanting to just have foundational training, or if they're spanning the available resources out there for AI? What are some of the resources you'd recommend?
Catherine Ordun:
Yeah, I think there's a full spectrum of resources. If you're just starting out and you want to get into some of the mechanics of AI, there's so many proven great resources like Coursera, there's DeepLearning.AI from Andrew Ng, who's a pioneer in this area. Google, Hugging Face, a lot of the big tech companies are putting out their own free tutorials and materials, and so I would definitely check that out. Much of this is self-paced, so you've got the luxury of doing it at home on your own time. For those who are more involved as a mentor and a subject matter expert, the Mark Cuban Foundation has been doing an AI bootcamp for a very long time, and it's not only for the teenagers who want to sign up, because they usually open it around springtime, but if you're a mentor or a subject matter expert, someone who's just really passionate about AI, and you'd like to volunteer and help these students, then there's also opportunities to connect you with Mark Cuban Foundation, and you can help them as well. So that's something that I think was really close to the heart.
And then also, for those folks who are just interested in getting more exposure to AI, funny enough, I would say that if you follow Twitter and there is the AI Twitter, there's a lot of discourse on that platform, and you can always find someone in your niche, whether it's art, economics, social theory, or just coding in general. Lots of personalities. For our Booz Allen folks, we've got the AI aware training, and that will really take you from soup to nuts, from just let's all level up on the same jargon, what's the same syntax, to a little bit more advanced material. And it's very well paced and easy to use, a lot of videos, and it follows common practices like you would see in Coursera.
Jennie Brooks:
Thanks, Catherine. Catherine, at the end of every podcast, we give our guests some free space to share their final thoughts. What would you like to leave with our audience today?
Catherine Ordun:
I would like to say that we are living in really exciting times. This is a once-in-a-lifetime technological revolution that we are seeing. I really strongly believe that a lot of the elements for generative AI are going to help us in the future just conduct our daily tasks. But what it really requires so it doesn't go off the rails, because a lot of controversy and concern about this existential dread, is that we as a community, as a society, all understand how AI will work for us, and how we can actually shape the future of AI. And this is not only a majority population kind of vote, this requires a lot of diversity from a lot of women, a lot of minorities to really get into the conversation.
Jennie Brooks:
Thanks, Catherine.
Catherine Ordun:
My pleasure.
Jennie Brooks:
Thanks for listening. Visit careers.boozallen.com to learn how you can be unstoppable with Booz Allen. Be the future. Work with us. The world can't wait.