This is Stare Down the Bull. I'm your host, Susan Hunt. Around here, we tackle the hard stuff, how to lead with clarity, leveraging AI, and turn strategy into real results. My guest today brings a powerful perspective on how to do exactly that. Let's dive in. Today, I'm joined by Dr. Judy Spitz, a pioneer in AI and speech recognition, former CIO at Verizon and founder of Breakthrough Tech, an initiative helping launch careers in computing. We'll talk about Judy's journey from early speech AI labs to the executive suite, what she learned about leadership and innovation, and how she stared down the bull throughout her entire career. Welcome, Judy. Thank you for joining us. It's my great pleasure being here with you, Susan. And thanks for inviting me. Yeah, I always love to chat with Judy Spitz. That's always a good time and a lot of great information. So I'll start out by just asking you, what drew you to AI and the speech labs so early on? I mean, it wasn't even in its infancy when you thought that this was a good idea. So how did you get there? Yeah, it's a great question. And it's one of the aspects of my career that I just consider myself to have been incredibly lucky with respect to, maybe being at the right time in the right place and so on. But I actually got my PhD in a field called speech and hearing sciences. And the best way to describe that is we were focused on trying to understand how we as humans came to understand and produce spoken language. and think of it like a branch of cognitive science. Visual perception, this was speech perception. And I was incredibly fascinated by it. I still think it's one of the most extraordinary and interesting aspects of cognitive science, how the brain works. And it just so happens that at that time, in that era of the relative early days of applied AI, people were asking the question, how can we get computers to understand spoken language? In fact, that might have been one of the first tasks that was defined as AI in its infancy, because AI itself is defined as getting computers to do something that we think of as uniquely human. Right. And certainly understanding and producing spoken language is something we think of as uniquely human. And so back in those days, the field of AI went about solving that problem by assuming that if we understood how the human brain did that function, then we could get computers, we could program computers to mimic that behavior. And so lucky for me, as I was being an academic and doing research in this area, a little known phone company called 9X, which is now known to the world as Verizon, created an artificial intelligence lab in their science and technology division. And who did they go and hire in order to work on this problem of speech recognition? but a bunch of people from academia who understood how humans did this. And so I simply moved a few degrees in a different direction from understanding how humans could do that task to saying, how can we get computers to do it? How can we take what we know about human spoken language and get computers to do the same thing. It was, and I think arguably still is, the most fascinating aspect of AI because fast forward, I'm either proud or horrified to say 40 years. What are we looking at as the most revolutionary aspect of AI, but something called LLMs, for those folks who don't know what that acronym stands for, it's large language models, which are really decades later, how far the field has come from understanding individual words to having a model of language that's so large that it can actually use that model to do much more sophisticated cognitive tasks than just understand what somebody's saying. But it is an evolution of the beginning was that started with speech and language. So I got really lucky that the field I studied had this application in AI and then I've been able to ride that train, if you will, all these years. Yeah. How did you feel working in those labs in the early days? I would imagine there weren't a ton of women working in those labs, but what was that like? And did you have any idea that it would blow up into what it is today and what it might become? Yeah, it was an incredibly exciting time for a couple of reasons. One was we were in the early stages of this fascinating field called AI and we all knew it. I mean, I can jump to your last question. No, I don't think people entirely understood where this was going. And with that, The flip side of that is I don't think any of us would have expected it took that long. Yeah. Because it's been many decades. Yeah. But there were some early futurists, most notably someone named Ray Kurzweil, who your listeners, if they don't know him, can go look him up. He's been predicting everything that we're seeing now since way back then. Yeah. So we knew we were doing something exciting. Think about it. The artificial intelligence lab that I joined, which was the beginning of my career, had in it a speech technology group, the one that I joined. It had a group called Machine Vision. And that group's first task was, can we read handwritten checks? Oh, yeah. That the phone company has to process. Because in those days, people paid their bills by writing checks. It had a group called an expert systems group. And again, for those of you folks out there who do or don't know the origins of AI, expert systems was really the beginning of what we now call machine learning. Because it was based on the fact that the way to get computers to reason through complicated tasks was to interview an expert. Yeah. An expert at anything. An expert in network planning. An expert in culinary design. It doesn't matter what it is. and extract out of them the rules that they use in order to be the experts that they were, and then simply program the computer to execute those rules. That has turned into, just think about it, those rules are still there. It's just that the machine is learning them on their own instead of interviewing experts. And then you have something called supervised learning, which is helped by experts and so on and so forth. But way back then, we had nascent groups looking at the parts of AI that really were a roadmap to where we are now. Amazing. Early on, did you have any stare-down-the-bull moments where you really had to overcome something that just looked so monstrous and so challenging? And if so, what was it, and how did you overcome it? Yeah, that's a great question. And I do talk about things like that oftentimes when I talk to young leaders because everything's not going to go smoothly. As I like to say, your career, you need to think of it as a jungle gym, not a ladder. So my career at Verizon lasted for 30 years and it was a steady but a slow climb. And so there were lots of bumps in the road. One thing that stands out to me, as I recall, because it felt like a career-ending moment, although it wasn't, we were trying to get the executive team, this is a few years into my career, we were trying to get the executive team to think out of the box in terms of internet enabling, getting the business online. The terminology we used back then is almost comical. We called it e-business. It basically means a digital corporation, everything online, but we called it e-business and moving the company onto the internet It seems very quaint now But those ideas were big transformational ideas back then And in order to get our senior executive team to think out of the box I brought in an outside, very, very edgy consulting company to do a workshop with some of our senior executives that would really put them out of their comfort zone. How did that go? When you're early in your career, deciding that you want to be the champion of putting your senior executive team out of their comfort zone, maybe wasn't the smartest thing to do. And it didn't go well. Let's just say that. And I took an entire day out of this team's schedule and they played along with me, but it did not go well. And I thought, well, it can't get any worse than this. Talk about failing publicly and colossally. I wrote emails apologizing for wasting their time. And the whole day was torture. I knew it was going badly. I learned a lot. By doing that, obviously, I survived it just fine. But I had another version of this where I got promoted in a pretty significant way, you know, a few years later. And something was happening with the organization. And I didn't like the way the decision making was happening. And I decided, since I was the new leader of this organization, I was going to write an email to a whole bunch of senior executives and tell them, very frankly, what I thought en masse in a group about the decision making was political and so on and so forth. I'm sure these folks had never received anything like this in an email. And very fortunately, my boss who had promoted me to this position, he ran defense with every one of the people who received that email and basically said to them and then said to me, rookie mistake. And so you're going to make rookie. And I have never forgotten that because it was a rookie mistake and he could have thrown me under the bus easily. And he didn't. He was like, yeah, okay, rookie mistake. So things will happen. We're all rookies at various times. And you have to be able to weather those storms and learn from this and add it to your bag of tricks as you go forward. I'm just going to say that I'm sure both of those groups of people, even though they weren't happy with your behavior, we're forced into thinking about what your points were outside of public display, because being that bold and taking those shots, I think is really important. And even if it doesn't turn out the way you think it should, I think it's far more important to at least take them instead of never taking them and just being a law flower. I got in trouble all the time at work because I was always taking those type of chances. And I say to people that I mentor, the amount of times I got in trouble compared to the amount of times that it actually worked out. It really worked out a lot more than I got in trouble. So yes, it was worth it. It's true. To make those kind of mistakes and take those kind of chances. Right. It's also a lesson, Susan, for us as leaders, as we see people coming up behind us, that you have to have your young leaders' backs. Yes. You have to be willing to not throw them under the bus, the opposite. To use them as coaching moments to defend them out in the world because they are going to make mistakes. So it was a lesson on both sides of the equation, if you will. Yeah. And when you do that, by the way, you do that. I always did that with my people. What it creates is a really loyal base of people working for you. They know they can take the chances that they need to take because their boss has their back. Even if their boss is furious with them for what they did, they have their back. And that gives people, I think, a lot more latitude to take chances to take bolder steps because they know. And those bold steps are the thing that make great leaders and great employees. I mean, you made those, you bumbled the ball a couple of times there, I guess, but you did become the CIO of Verizon. And you also did happen to deploy maybe the first largest call center speech recognition ever in the country, maybe in the world. I don't know if it was the biggest, biggest, but I think it was. Well, I did that. And you know very well that you and I did that together. Well, I sold it to you and you did all the work. I got yelled at occasionally. But no, you did that with your team and it was really inspirational. Amazing. Yeah. It really was. Do you want to talk a little bit? Could you talk a little bit more about your time as CIO and how that transpired? Because while you are talking about some things you did early on in your career that were wrong, you did so many things that were really right. And you were so well respected at Horizon by everybody, every executive and the people that work for you and with you. So tell us a little bit more about how that transitioned over the period of time you were there. Yeah. Thanks for the compliment. nice to hear my transition from member of technical staff to cio took 20 years i mean i'm proud of it but i wasn't on a rocket ship and i think that people need to understand that you have to put in the time and prove yourself so one of the things i think as i look over that career arc is number one as i like to say to people start doing the job that you want to be promoted to because you have to give people a vision of what you're capable of beyond your just asserting that you can do more than your current job description. Yeah, good advice. That's really good advice for younger people. That means taking risks. And so there were several times where there was an opportunity to raise my hand and say, I'll take on that risky project or I'll take on that thing that allowed people to see me in a role beyond what my current job was. And so that's sort of one observation. And in each case, it led to a promotion. In every single case that I made a move like that, it led to the next promotion. And similarly, my actual promotion to CIO came out of the same thing, which was I was the senior VP responsible for IT for a particular division of the company. And I just had done some great work with my team in delivering to that organization. And then we merged with MCI for those folks who know the telecommunications history. And it just so happens that one of the senior leaders of the team I was supporting went over and became the president of the new business unit and pulled me over and said, I want her to be my CIO. And that came directly out of his experience with knowing and trusting my ability to lead and deliver and so on. So that was that last piece of that move. So the lesson that I like to share with folks is, number one, start doing the job you want to be promoted to. And number two, you have to do a great job every day. Yeah. Not just on special days, not just on big presentations, but every day you got to come to work and you have to do a great job because people will remember that when they're looking at who do they want to bring with them. And the other aspect of it is, I think it's fair to say I was a non-traditional CIO in terms of my background. I think, Susan, you know, I don't have any computer science degrees. I don't have any IT degrees. I have a very strong foundation of science and technology and systems thinking and so on and so forth, but not traditional. And I always understood that. And this is a leadership lesson writ large. Never tried to either cover that up or pretend I was something that I wasn't, but rather understood the value that I brought because of my orientation and my leadership and my systems thinking and so on. and then surrounding myself by people who were much smarter than I was in the areas where I had less expertise. And that both willingness and, in fact, passion around making sure that the people who I surrounded myself with in my organizations always were smarter than I was and knew more than I did allowed me to leverage my strengths and leverage their strengths and also give the view to the broader organization that collectively we could deliver. And it is a style of leadership. It is. It definitely is. That worked for me both on a personal level in terms of what felt authentic to me, but also worked for me because of the nature of who I was in the organization. It was really the only way for me to be successful was to step into that cycle of leadership. I think the work that you did with regards to deploying the speech recognition in the call center was very unique in the way that you did it because of your understanding of the customer and how that would work along with your experience in speech and all of that. So that was really, really well done. And while you may not have had the IT background, you had a background that most CTOs that also deployed technology in their companies did not have. And that is a clear understanding of how people think and how speech recognition actually works and why it's going to be valuable in a certain way rather than a different way. So, yeah, that's true. And if you fast forward, probably the largest program I ever ran at Verizon, short of just running the organization, was the Fios deployment. So as you may remember, number one, that was a multi-billion dollar bet company program at Verizon. And I was responsible for overseeing all of the systems support. There was an engineering organization that did the build. There was a marketing organization, the product organization. And then there was the systems built. And there again, I would highlight the two things that allowed me to be successful. One was, of course, surrounding myself with all of the IT experts that knew how to build those systems. And the other was being able to take the customer's perspective. So in the case of the voice portal, as we called it in those days, yes, I did have the subject matter expertise about communication. But it's also just a general orientation around understanding the bigger picture around the business and understanding what the customer experience is and means to be. and that is a value proposition that has always held me in good stead and I think most people would agree that getting somebody in a technology field also understands the business who also understands the end customer is a value proposition that often differentiates really great tech leaders from not as great tech leaders. I couldn't agree more. I really true. So after you did all that, you decided to start up your own little thing, we'll call it, at Cornell. I'd love you to tell us about Breakthrough Tech and why you decided to do that, how you went about doing that. It's really incredible, the work that you've done over there, too. Yeah, this is my, you can call it my passion project or my second career, but after having been at Verizon for 30 years and having been the CIO for one of their business units for the last 10 years, it was time to do something else. And anybody who knows me from my Verizon days knows that I've always been passionate about bringing other women along, especially in tech careers. As I like to say it, I think we all have an obligation and I have a passion for lifting as you climb. And so even while I was at Verizon, I think I had a pretty strong reputation for supporting the other women around me and making the observation about how few of them there were. And so as I thought about what to do next, I really just sort of grappled with the, holy smokes, what's going on here? I've been here for 30 years and there seem to be not very many more women in the tech career path than there were 30 years ago. And I just thought to myself, why don't you take 30 years of management experience and try to point it at this problem? And so my goal was less to have a little hobby and a nonprofit than it was to really try to look at the system. What were the systems that were perpetuating this extraordinary gender gap in tech? And those systems include, of course, higher education, because that's the feeder pool, that's the supply side, and the industry hiring engines as the other side of that supply chain, and all of the white space in between. So here are a couple of factoids that are still true today. 58% of the undergraduate population are women, and only 1% to 2% of them study anything related to tech. That's just crazy. It is. It is. So on the supply side, you have to scratch your head and say, what's going on here? On the demand side, the observation is that industry was then and continues to look for and recruit tech talent the way they've always looked for and recruited tech talent, which is they look for certain things on resumes and there's a feeding frenzy looking for talent at a very small number of universities basically the top 20 or 25 universities well guess what turns out that 80 of the women and 90 of the black and latina women who are studying tech don't go to those top 20 schools they go to 1900 other schools around the nation and they've got all the potential in the world to contribute and enter and thrive in the tech industry. What they don't have is the privileged access to the opportunities that their more privileged peers who go to those top 20 schools have to get the things on their resumes that would allow them to get their foot in the door. And having been in corporate America all those years, I don't think that it's being bad corporate America or anything like that, but you can't recruit from a thousand different schools. So you have to narrow the focus. And if the only tool you have in the tool shed is a summer internship in order to find and identify talent, and there aren't enough of those to go around, you'll never see that talent. And I'd like to think about this. Susan, you'll appreciate this. I use the example of call centers that we're both so familiar with. It has always been the case for technology changes that you can't take a new technology and bolt it onto an old system and expect to see transformational results. So yeah, you can take a speech recognizer or today's AI engines and so forth in order to try and transform a call center. But what doesn't work is to bolt it on the outside of the call center, do the best you can, and then when you fail, throw it over the wall to the old process. That doesn't work. What you have to do is re-imagine how you're running your call center from the ground up in order to take advantage of something like AI and spoken language systems. And there are other examples around cloud computing and cloud-based software development. There are examples around mobile app development. I have simply observed that the same thing is true in tech recruiting. Now, you can't assert that you want to find a different and more talent and yet continue to recruit exactly the same way you always have. And so when I launched Breakthrough Tech, it was as a, what we call a nonprofit intermediary, fancy way of saying we sit in the white space. Yeah. Between industry and academia, and we try to create the kind of programs that would simply let these students show what they're capable of. Because if you do that, the rest takes care of itself. I will note that while when we launched, we were exclusively focused on women, that is no longer the case. We are an inclusive organization and we're mostly looking at talent that goes to universities that are more often than not simply overlooked by the hiring engines of the tech industry and trying to solve that problem which is how do you take this overlooked talent and give them access to the kinds of experiences that will allow industry to see their talent And it means bringing both sides of the supply chain to the middle You have to get industry to lean in in more what we call work-based learning experiences. You have to get higher ed to lean in in terms of creating the space and assigning value to experiential learning and work-based learning experiences, and the results are pretty phenomenal. I think you probably know we run now the largest AI talent accelerator in the country with well over a thousand students each year. It's amazing. And they're coming from all across the country. This is overlooked talent that is landing jobs right now at a rate of over 80 percent, and that's in today's market. So it's amazing. We know we can solve this problem, but there's more work to do. Congratulations. Yeah, for any of those companies out there that are looking to get some interns from your program, we'll make sure that they have a way to contact your organization. Absolutely. Because they're always looking for more places with these very smart young people. And I've been in three companies now where we've had their internships and they've done phenomenal work. So really phenomenal. Yeah. So I want to talk to you a little bit about AI ethics and bias and maybe the future, because it goes back to that few women in tech has now created an environment that we have a reasonable problem with AI bias. if all engineering is done by men or mostly by men, there's a lot of bias in AI that I think is worth a conversation and talk about. What do we do outside of creating jobs for women in tech that can also, which I'm seeing, by the way, I'm seeing women coming into the AI space more than ever before, which is interesting. So what's your take on AI bias and what we can do more than we're already doing and what you think the future holds with that. I think it's an incredibly important question. I'll simply make the observation that this concern about bias and how it results in products and services actually privates AI. So as your listeners may know, the early design of airbags for cars actually injured a lot of women and children. Why? Because the dummies that they used in the test labs of the automobile factories were all based on the size and structure of the men in the labs. And therefore, the airbag interacts with the person sitting in the seat very differently if that size and structure is the woman or a child. And that's long before anything related to computers or AI. Fast forward, you know that the early speech recognition systems did a lousy job of recognizing the female voice, also a lousy job of recognizing accented speech and so on. And all of it comes down to what I describe as the who, meaning it matters who's in the room, who is in the room when these systems are being designed and built and trained and so on. because when it comes to something like algorithmic bias, for the most part, that can be traced back to what's the training data. And the question is, who is in the room when the training data sets are selected and being used and asking the hard questions? Who's in there asking whether the data is representative of the population? You know, I'll give you examples that go well beyond gender. again, you probably are familiar that facial recognition systems have been notorious in the last five years for leading to the false arrest of innocent black men and women because these facial recognition systems do a significantly poorer job of recognizing the faces of people of color. There are systems that have been trained that have to do with, for example, where police cars troll and don't troll that are biased by the demographics of zip codes and so on and so forth. So this question of bias and training data extends far and wide. And it also goes to beyond just the bias of the algorithm to questions of fairness. Is a false positive the same penalty as a false negative? And for who? Is the outcome of that decision fair? And who is it fair for? And I believe that all of these things we can worry about and we should worry about policy, worry about governance, worry about regulation. We should worry about and we should do important work in each of those areas. But I think the first and most important determinant about whether this extraordinary AI technology is going to serve overall to uplift or endanger society is not going to be based on any of that. The first principle, in my opinion, is going to be based on the who. Who is in the room? Who's deciding, well, which product should we invest in? Who's deciding what the training data is? Who's asking the questions about policy, about deployment and so on. And that's why I think it's almost more important than it's ever been before to make sure that the people who are in the room designing all of our collective futures are representative in all respects of the society that these systems are going to impact. No. And you know for sure they're not right now. That's just not happening, I don't think. I mean, There may be smatterings of people representing parts of the world, but certainly not the entire representation. I think that's what's a little nerve wracking. Yeah. What we do know is that it's not going to happen without intentionality. Yeah. That right now, the decisions are being made by a very small number of companies and a very small number of people. And even if you widen the aperture to the people who are in the rooms designing and developing and so on, it is a very narrow slice of society. We're going to have to be intentional about who gets access to the kind of training. And you can go all the way from what I call a light lift of AI literacy. Do you have even a sense of understanding how these systems can and should be used such that when you get into a job, you can be in those rules and you can be knowledgeable enough. You don't have to be writing the algorithms, but you can be knowledgeable enough to ask the foreign questions all the way down to the people who are actually developing the algorithms. You know, again, your listeners may or may not know every single algorithm that drives one of these systems is being optimized for something. It can be optimized to minimize false positives. It can be optimized to drive positive. It's being optimized for something. And the question is, who's in the room asking the question about what the algorithm is being optimized and challenging whatever that assumption is. So I think there is a, what I like to say is the fierce urgency of now to make sure that we are being intentional in creating the opportunities for anyone and anyone across all demographics, all socioeconomics, to get the knowledge base that they need to get into those routes. Yeah, that's awesome. The work that you're doing is really important, Judy. Thank you so much for putting together Breakthrough Tech and creating such a large, hopeful environment to solve some of these big problems. I mean, we can just wish it goes away or we can do what you're doing, which is actually creating AI future leaders to help with the situation, with the bias situation. I really appreciate you being on Stared on the Bull today. It was a great conversation and looking forward to seeing you again really soon. Thank you so much. Thanks for having me, Susan. It's always great having a conversation with you. Thank you. Thank you for listening to Stare Down the Ball. If today's conversations spark new ideas, share it with a colleague and keep the momentum going. The future belongs to leaders who act, so subscribe and join us again next week.
