Andrea Gallego Global GAMMA CTO, Managing Director, and Partner of Boston Consulting Group (BCG)
Throughout my career as an unwavering advocate of AI and AI for good, I’ve had the pleasure of meeting, connecting with, and absorbing wisdom from a variety of individuals. As part of my 10-Part Series of The 9 Inspirational Women Leaders In AI Shaping The 21st Century, I was honored to converse with Andrea Gallego, managing director, partner, and the Global GAMMA CTO at Boston Consulting Group. Andrea is a thought leader in artificial intelligence (AI) and its impact on business and society.
She is a significant proponent of Women in AI and has been working to increase the visibility of women in the field. Throughout this interview, I learned about some of the most interesting trends in AI that Andrea is observing and what she believes will have the most significant impact on business and society in the years to come. Specifically, we discussed: the importance of diversity in AI, the potential for digital twins to transform manufacturing and supply chains, the future of quantum computing, and the need for policies to ensure that AI is ethically sound. I am delighted to share this conversation with you in the hopes that it may inspire you to think more deeply about the impact of AI on our world and awaken masses of female talent in the field.
BRAZIL – 2019/07/11: In this photo illustration a Boston Consulting Group (BCG) logo seen displayed … [+]
Sure, of course. I was a full-year associate at Lehman Brothers for about two months during the crisis before the Lehman Brothers Manhattan office closed down.
Afterward, I worked for a private foundation that a mathematician and his wife run. The husband runs a hedge fund. They wanted to use their money helpfully and discover great causes. One part of my job was to find out how we could be sure that the money we were giving was put to good use by the institutions that received it. I learned that in the realm of non-profit, it’s rather emotional and focused on relationships, and it’s based on a very tight-knit circle of trust. However, it was supported by data. There was little rigor around how many cents on the dollar went to the programs versus the administration of the NGO you were donating to or the organization itself (what are the mechanisms of reporting, etc.). So I recognized that I wanted to do good in the world but that it needed evidence and could be quantified and backed by data.
So I went to grad school for my computer science and analytics degree, where I blossomed and became interested in analytics and AI. However, I was always interested in being able to scale it. I was more concerned with carrying out plans than with math. If you make this remarkable algorithm, how does it make its way into the world? How do you operationalize and scale it? This is what I was interested in pursuing. And so I started my work at Booz Allen, and then my work at Mckinsey, continuing to scale data science. I’ve been doing this for over a decade now, and when BCG called, they were like: “We’ve been doing data science for a while, but now we need to scale it. Will you be our CTO?
And of course, I thought this was precisely that combination of technology, AI, and scale that I was looking for, and it was fantastic timing, so I joined BCG.
More than ever before, I believe that it is essential to look at the data released regarding women in STEM and what’s happened since Covid. We’ve regressed regarding women’s rights and women in the workplace. We’ve lost ten years of many of our female leaders’ efforts in the 1960s, 1970s, and 1980s.
We still have our right to vote, but beyond that, we’re going backward and not forwards. I believe that part of my job as a leader and other female leaders is to ensure that we give our younger women a voice, that we demonstrate to them that we’re not only staying in the positions we’ve got but also climbing up the ranks.
We need to show other women that we can make it work and that the organizations need to help us make it work for us, whether we’re moms or caregivers or anything else.
This quote was inspired by one of my favorite leaders – Steve Jobs. There is a book that Walter Isaacson, a fantastic writer, wrote about Steve Jobs, and one of the lessons I took away was how eager Steve was for anything non-technological: arts, calligraphy, design, and writing. That’s what made him such a fantastic technology leader. I think we’ve lost the desire to be interdisciplinary and have too much focus on perfection in one field – the feeling of I need to be a computer scientist, or I need to be a data scientist, or I need to be a mathematician, or I need to be a chemist. I have to have a Ph.D.
We’ve lost sight of what makes technology beautiful?
What makes technology beautiful is being creative, designing art, and making it accessible and usable to the entire population.
And that takes much more than just an engineer, you know? It takes photographers and designers, and I think it’s important for women to understand that you don’t have to be just a computer scientist or a data scientist to be in this field.
I believe there is a disconnect between education and the arts. I’ve made the case that today’s educational system has done a disservice by separating art and science. You must pick one or the other when you enroll at many schools and earn multi-disciplinary degrees in some niche specialties. We need more of this interdisciplinary research.
Sociology and anthropology, understanding how people respond and react to technology, will be critical as we commercialize and make AI a commodity in the world. And so the question is no longer: “How fast is my algorithm?” The question is that when intelligent machines run someone’s house, how will they live? What is that going to mean for a 20-year-old versus an 80-year-old? And, what is that going to tell a man versus a woman?
All these things matter, and so that study of persona, and that study of humans, in my opinion, needs to come back to the forefront.
I think first we have to be very honest with ourselves. As a species, we’re inherently biased. We’re biased, not necessarily in a negative connotation, but our upbringings limit us. We all have a different upbringing, different parents, and come from different countries, which creates unconscious biases, whether we like it or not.
We will not be able to fix that overnight, which means all of the data that the machines produce about humanity are inherently biased, and the first step is to acknowledge that.
We need to acknowledge that the data is biased. Remember that when you start looking at that data, we need to understand first: “What is the bias I’m looking at? Is this data leaning one way, or is this data leaning the other way? What is the algorithm that I’m building? What is its intention? “
I always use the example of building hospitals across a nation. If you were to simply input the algorithm into a model, it would optimize for-profit and minimize cost. So, in other words, you’d place the hospital in an affluent location where you could pay your doctors the best wages, and no hospital would be built in a poor neighborhood where a need truly exists.
I wish this weren’t the case, but I don’t think we’ll avoid regulation.
First of all, I believe CEOs are concerned and asking the right questions about AI and its impacts. However, there’s only so much one company can consider at once; They are busy ensuring their company stays alive so they employ individuals and remain competitive as a business, which implies they may be short-sighted in what will happen in ten years if an algorithm somehow had malware or impacted a population the wrong way. And so sometimes you need an unbiased body of people to remind you of certain things that you’re blindsided by. But as many know, regulatory authorities need to be formed by the right people and have the proper governance to be helpful and further innovation vs. stalling innovation. We’ll need to ensure we do this right.
The two things that scare me the most are bio and cyber warfare.
You might be able to construct some devastating viruses when you combine AI with biology, healthcare, etc. Then, when you consider cyber warfare, the electric grid comes to mind. If someone hacks into that grid in the wrong way, we could lose power and never get it back.
That is a very, very dangerous situation. I’m concerned less about the old War than I am about the new war tactics developed.
One of the significant challenges we have is: first of all, figuring out the data. We were in a position where data flow was crucial, correct? To mine and gain access to all of the data, we needed to be able to store it all in one place. Now that we’re facing many geopolitical issues with data, this is changing.
It’s a huge obstacle. If you’re a global organization, you need access to global data and flows of that data to understand how your supply chain is impacted globally. So we’re all facing this and trying to figure out how we share data as a world? And one of the great things that came from sharing data, as you can see, is the research that came in developing the COVID vaccine, where many data points across the world were used. However, it shouldn’t be necessary to have a worldwide crisis to realize that we need to share resources and data to advance. So I’d say that’s the first one.
The second one is that it is challenging to modernize legacy systems. It’s challenging, and it takes a lot of investment, so there are some tough decisions that leaders have to make: “Do we just sort of cut ties with what we have on the legacy side and create something new? Do we try to modernize our legacy stack slowly and move fast? This is a difficult conversation, and It’s a long journey. Many people believe that they can just digitally change overnight, and that isn’t the case. It takes time and resources.
The final barrier is that many people want to get started quickly and on a modest investment. They observe another firm as a digital native, and it’s one of those instances when you come upon an entrepreneur who has succeeded. “I’m just going to be an entrepreneur and be as successful as they are,” they think. What you’re missing out on is that there’s a lot of years of effort behind it. So I think there’s some truth and practicality in this. I believe specific measures must be taken and consistency and discipline. Knowing that the process won’t take six months is essential.
One of the projects we’re currently working on is natural language processing, which aims to assist quality engineers in addressing medical technology complaints. I think that’s an excellent one because it may help make medical technology better in a world where there are many problems. Even if the FDA approves these technologies, complaints provide essential information to improve or even make them safer.
It’s challenging for a person to go through all of these complaints and be perfect. And so, the ability to utilize natural language processing to assist humans in comprehending them is something you should examine. These are the essential complaints because they’re the only ones a computer can judge with little effort.
I believe it’s inspiring, and I think it has a lot of promise for improving health care in general.
Look, I’m willing to bet that there will never be a match between a machine and a human as long as people are involved. Humans will always need to be part of the equation, and there will always be flaws in the technology; for one particular reason – the machine is logical, and the human is emotional.
What do we mean when we talk about this in the context of self-driving cars or having a flawless conversation with an AI? The machine has no idea how to react to an unreasonable argument. And as human beings, we engage in irrational discussions at one point or another. They’re based on emotional drivers that the computer can’t interpret.
So I believe that over the next decade, yes, some jobs will be lost, and this is where I propose our educational system should receive priority in addition to our climate so we can prepare to reskill people. Still, I don’t think it will happen in the manner of “machines taking over the world.”
I believe there are several things that humans shouldn’t do anymore across all sectors – mundane things that humans don’t want to be doing. And that is also very error-prone because they are very repetitive, and that means the human brain just sort of makes mistakes because you know how the human brain works – it tries to fill in the gaps. So we can make some mistakes when a machine is ideally intended for right. This is where the enterprise can figure out exactly where to place AI in that sector. And this is where I believe we’ll see the most utility, as well as automating those areas and then finding higher-level functions for humans to execute.
I believe that the first step a woman should take is to look for employment in fields other than computer science or data science. Look at the whole value chain, consumer, product, and stem; examine all parts that influence artificial intelligence. Find someone with the same mindset, interests, and goals. Find that person, find that network, and interview many women. There are women at the MIT Media Lab and Yale labs doing incredible work in creative arts and understanding the brain using data visualization. In other words, expand – don’t think that the language is just computer science and data science. Do some research and find women and speak to them.
Take a Udemy course, take a Coursera course, and look at what’s available. Look for job descriptions that aren’t just data science or computer science.
I think what Girls Who Code is doing could not be mentioned more. I think what they’re doing to specifically find underserved areas where girls are not having those opportunities and putting them in Stem programs is one of the best things we could do.
The other issue is that we need more women to be teachers. We need more women to be teachers so women can learn from women earlier on. Teacher salaries, in general, do not provide for such an incentive; as a result, there are few female (and male) math and science instructors in the United States.
We need to figure out how to incentivize females in science and math to be mentors, teach, and even go to school part-time to say: “I can be like her one day.” We need women from various backgrounds who speak different languages to communicate with these young girls.
I’m interested to see where we go with hardware and battery power because of its impact on climate. And I think it’s coming close to the innovations I’ve seen come out of Intel and Nvidia. Some battery companies are super important and very interesting because as much as we want to avoid the situation, AI creates a lot of Co2.
Until we figure out how to get that power on a smaller chipset, with less power, the more we’re going to run into this situation of If we go beyond quantum with the power we have now, we’re going to be emitting a lot of Co2.
So I’m super keen on the fact that Intel, Nvidia, and the other companies seem to be creating chips that demand almost half the power that today’s chips require to run AI systems, and I’ve been watching that one closely.
I think so. Because when you think about whether or not we’re going to grow as a globalized economy or shrink, it’s going to be very important to be able to take supply chains and put them in different areas, and test them in other areas, without actually having to build the entire infrastructure.
When you consider the ability to produce the same thing in several locations and the opportunity to test it without going through the same time-consuming procedure, which also impacts everything, it affects the climate. It makes things move faster. It affects labor.
It will be a game-changer if we get it right, especially in manufacturing and the supply chain.
Throughout this article, we have discussed the importance of diversity in AI and some of how we can improve upon current initiatives. We have also looked at some of the most exciting trends emerging in this field and how they may impact the future development of AI. I hope that the private and public sectors can continue to work together to ensure that everyone has the opportunity to benefit from the extraordinary potential of this technology. Lastly, we must emphasize female involvement in AI and incentivize and encourage women to enter the field as mentors for the next generation of innovators. We can create an even more inclusive and equitable society with their unique perspective and insights.
Andrea is the Global GAMMA CTO, Managing Director, and Partner of BCG, focusing on AI at scale and building digital and analytics solutions across multiple industries. She Founded Gamma’s engineering team and Source AI, GAMMA’s first AI/ML software, which DataRobot acquired. As a female leader in the STEM field, she is a strong supporter and member of Girls Who Code, Women Who Code, the AnitaB.Org foundation, and helps run Women in Digital efforts at BCG.
Disclosure: Mark Minevich acts in capacity as external advisor to BCG