Episode 14 – Building Trust in Government with Responsible AI
Apr 19, 9:00:00 am
FortressIQ | Intelligent Insights for the Modern Enterprise
Episode 14 - Building Trust in Government with Responsible AI
A framework for AI that leads to successful outcomes includes proper governance, thoughtfully conceived processes based on input from affected stakeholders, and transparency about AI’s role in decision making. Responsible AI represents this comprehensive approach. To achieve it, governments must empower leadership and use AI to enhance human decision making—not replace it. A Responsible AI approach should include regular reviews, integration with standard tools and data models, and a plan for potential lapses.
In today’s episode, host Jon Knisley, long-time technologist helping companies win the market with emerging digital solutions, and series producer Elizabeth Mitelman talk with Kathy McNeill the director of artificial intelligence at the GSA’s Technology Transformation Center of Excellence. Kathy dives into the role of proper governance in AI, automation, and technology in order for tech to remain true to the desired outcome it was originally designed for. She also talks about the challenges she has seen in the industry and her advice for women who want to have a tech career. Kathy also shares why passion and being technically-curious is important and how to be a participant and a life learner in the industry.
The critical role that GSA plays in the government’s overall operations in technology leadership
Kathy’s role as the the director of artificial intelligence at the GSA’s Technology Transformation Center of Excellence
Kathy’s biggest challenges and lessons from the industry and about how the government is changing and adapting to different constituencies
Advice for women interested in a career in technology, in the public sector
Kathy’s journey into the tech industry
The current state of AI adoption and governance
“The Missing Middle” and how to close the gap
How to ensure that programs have the right checks and balances in place for responsible AI ethics
If you enjoyed this episode, subscribe and check out our series at fortressiq.com/podcast. Thanks for joining us today on hello, Human.
Full Episode Transcript:
John: Hi and welcome to hello, Human, a podcast to explore ideas and feature humans working in AI and technology.
Kathy McNeill, the Director of Artificial Intelligence at the GSA’s Technology Transformations Center of Excellence, joins us today on the hello, Human podcast where we discuss the latest topics in artificial intelligence and how it’s being applied in the real world.
I’m John Knisley, the host of hello, Human, and a long-time technologist helping companies adapt and utilize emerging digital solutions. A big thanks to FortressIQ for sponsoring the program. Be sure to subscribe wherever you listen to podcasts.
This episode is part of our special series on Women in AI that we are very excited about here at FortressIQ. Elizabeth Mitelman from our marketing team who has been a key driver of this special series is participating in this session as well.
In this episode, we are going to explore building trust in government with responsible AI. Even when you take the politics out of the last few years of the equation—and I promise that we are not going to go there—trusting the government has been at a historic low since all the way back to the 1980s, according to Pew Research.
Then you add in the technology and AI on top of an already skeptical population. It becomes an even more interesting issue to explore and I’m sure more challenging to address. To help achieve trusting government with responsible and ethical AI, governments must empower leadership and use AI to enhance human decision-making, not replace it. This promises to be a fascinating discussion.
Welcome to the program, Kathy. Thanks for joining us on the hello, Human podcast and participating in our Women in AI series.
First off, as you’re aware, I spent some time not too long ago in the public sector and actually supported the GSA’s program, as well as the Defense Department’s Joint AI Center. But many listeners may not even realize that GSA stands for General Services Administration, let alone that it employs close to 12,000 people and has an operating budget of more than $20 billion. Obviously, a large and complex enterprise to say the least.
Maybe you could start us off with just some insights on the critical role that GSA plays in the government’s overall operations in technology leadership, as well as your role in the Technology Transformation CoE leading the AI area.
Kathy: Sure, John. Thank you so much for having me and thank you to Elizabeth as well. I’m really honored to be part of this really distinguished group of women in the Women in AI series.
General Services Administration, GSA services, have a really important role in the government. We support agencies as they procure technology, as well as other services like buildings and facilities. That’s what we do overall.
Within GSA, there are many branches. One of them is the Technology Transformation Services Group. It’s a relatively new group and it’s been around about four years. Its role in the organization is to provide transformative technology services to support agencies. Underneath that umbrella is the Centers of Excellence. I’m a director of artificial intelligence within the Centers of Excellence. Our mission is to support agencies as they are going through IT modernization and transformation.
We have multiple centers of practice. We have the infrastructure. We have cloud adoption. We have contact centers. We have customer experience and of course, my favorite is artificial intelligence along with data analytics.
How we do this is we partner with agencies. We help them define strategy. We help them look and procure the proper technology. We also partner with our industry experts to help our agencies also drive forth under strategic initiatives. That’s what the Center of Excellence does.
John: That’s great, Kathy. Thank you so much for that overview. I’m fairly active on LinkedIn and try to post once or twice a week. The post that has generated the most engagement in terms of comments, likes, and shares—whatever you wanna measure—was about the State of Federal RPA Report that noted 848,000 hours of savings last year alone from about 500 bots in production. It’s just an amazing number and one that I’m sure that many commercial enterprises are quite jealous of.
Speaking of the private sector, I know that you spent a fair amount of your career in the private sector as well. Can you talk about the transition to the public sector? What’s been your biggest surprise? What’s been the biggest challenge? Some of your learnings there.
Kathy: Sure. Here’s the thing, the government and large organizations in the private sector, there are a lot of similarities that they have and yet there are some differences because a lot of the activities that the government does are very unique to the government.
For example, the drivers in the government are different than the commercial sector. Government drivers are about serving the public good and are rooted in public policy. In the commercial sector, you’re dealing with market drivers. You’re dealing with revenue goals. That’s one of the differences there. Government services must be inclusive and support diverse populations across economic, geographic, cultural spectrums whether they be English speaking or not, rural or urban, with internet accessibility or without.
That’s super important as we go down through this process. It means that all the processes that we have within the government have to support and adapt to all of these different constituencies all the time and consider the needs of our country’s many diverse and underserved communities.
That’s a lot different than a commercial environment. Government systems themselves are very disparate. There are areas in the government where technology is very advanced with new technology and platforms. Think about the cloud, think about mobile, high automation. Other areas in the government are still very much working with mainframes and traditional on-premise systems. There’s a lot of disparity in technology.
One of the things that I found really interesting—I don’t know if it’s a surprise as much as it’s interesting—is that there are very sharp people working to make the government better. They believe in the missions.
For example, I want to use something that GSA did. The usa.gov’s contact center is under the GSA umbrella. The center had information to assuage the fears and confusion when COVID hit. In a matter of weeks, the contact center was moved off-site so agents can safely work from home. The usa.gov implemented an IVR, an interactive voice response system, that was essential to respond to the surge in questions.
There was a surge of 80,000 questions in March 2020 alone about COVID and the pandemic, 650,000 calls and chats since this pandemic began. The focus on self-service options for customers helped ease the burden of agents.
You made a reference to the bots there. The IVR actually set up some bots to help handle the calls and about 50% of the incoming calls were handled with automation. That’s pretty impressive from my perspective and something that I found surprising underneath the covers in our government.
The government also has a wide array of services—things like financial, military, human services, and medical services. What industry under its umbrella has all of those different characteristics? Even small agencies are larger than some of our mid-market commercial agencies. That’s really surprising to me. It’s a very big challenge when we’re talking about what the government is doing to serve the population.
John: That’s a great example. I love the example of the IVR system. I’ve lived and worked in the DC area for 25 years now or so, so I’ve had opportunities off and on in my career to be involved with public sector projects. I think that idea of mission that people feel who work in the government is so critical to the success. That idea of, I’m contributing to the common good for everyone I think is a pretty cool factor.
This episode is part of our Women in AI series. For our listeners earlier in their career, what lessons would you pass on for people interested in a technology career in the public sector?
Kathy: This is a great question, John. I reflect on my career and I’ve had lots of twists and turns. It’s not a straight progression. I’ve got an undergraduate degree in English and a minor in Math and Economics. I get a lot of questions about that. I have a master’s in computer science I got after working about eight years or so.
I started out as a tech writer many years ago. I was very fortunate, there’s no doubt. I had the aptitude for technology and I had some mentors along the way. They were men. It’s okay that we have male mentors and female mentors. Let’s take that into consideration as we’re going through our careers.
When I started, technology was exploding as an industry. How women would progress through those ranks is a bit different than it is today and yet it’s still the same in some ways. Moving into the public sector after spending time in the private industry has some advantages. I appreciate the complexity of problems that are being solved in the government. I don’t know that if I started out in government, I would have appreciated some of that. I understand how complex problems can be solved in the government. I bring my commercial sector experience with me.
Women entering into the public sector workforce have the opportunity to really make a difference in the agency they serve. If I were to advise somebody, I would say pick an agency where you can support their mission and have a passion for what they do. I chose to come to work for the TTS Centers of Excellence because their focus is on IT modernization and technology transformation.
In my core value, I believe that technology can better customer experience and can better how they work with the public constituency. Women also bring a very rich perspective to a technology team. We need to appreciate that and we need to recognize that in ourselves. In any organization, I feel it’s extremely important to be true to your values, as I said earlier.
As we are starting our career, find out the things that bring you passion, bring you joy, what kinds of solutions you can bring to the table. It’s also important to realize that we’re not going to get it right. There are many missteps we’re going to make along the way. What do you do with that? Learn from it. Learn from ourselves. Learn from the mistakes. Admit the mistakes. Learn from others. That will make us grow better in our careers and make us better as people and contributors to the public sector.
One of the things that I see as a challenge that we had 20 years ago and we still have today—I’m sure I’m not the first one to say it on your podcast, John—is getting our voices heard. Women need to speak up. We need to volunteer for the tough project. That means we got to raise our hands. We got to ask for the new position and we got to be tenacious about it. We don’t need to take no for an answer. We got to keep asking. Don’t take no for an answer. That’s really important. I think sometimes, we get a little too polite and we need to keep focused, keep going, push on.
Elizabeth: Being a young woman in tech, that’s really, really great advice. Being able to stand up for yourself and get those projects going is exactly how you do it.
Kathy, thank you again for joining us for today’s episode of hello, Human. Turning into your current role, how would you describe the current state of AI adoptions in the government? I think that there are a lot of people that would be surprised by how many agencies have active programs in place so I’d be curious to hear it.
Kathy: Thanks, Elizabeth. You’re right. There is a lot of artificial intelligence, machine learning solutions being implemented in the government. We see a lot of interest in AI adoption and a huge range of what agencies are doing and trying to do with artificial intelligence. It is being used as a tool in the toolbox to help modernization move forward.
The level of adoption, of course, varies by agency depending on its mission drivers. We see agencies involved with RPA, much like John mentioned earlier, with all of the bots being implemented. Machine learning natural language processing is a focus right now and many others.
Within GSA’s TTS organizations, we have something called the community of practice. We also have working groups with focus areas that the federal members can share and collaborate on. There are community practices for artificial intelligence, RPA, machine learning. What’s great about these forums is they’re open to anyone in the federal government, any level. They talk about all kinds of topics including implementing the technology, procurement of solutions, and assistance on specific technical issues.
For example, the AI machine learning community practices over 1000 federal members. I found that to be pretty amazing to see the level of adoption there. TTS CoE’s mission is to continuously support agency partners as they strategize, road map, select, and implement artificial intelligence solutions.
We talked about that a little bit early on, but just to reemphasize that. We accomplish that in a variety of different ways. We partner with agencies skilled in this to help us along this journey.
John: That’s great, Kathy. Digging into the adoption a little bit further, the term the missing middle often gets used to describe the AI use in the government. There are a lot of very advanced programs on one end of the spectrum. Then on the other side, there are a lot of emerging programs that are still in the lab or still being prototyped. Do you think there’s anything that can be done to really close that gap and drive more adoption, or is it just a matter of time and these programs need some nurturing just to allow them to grow and prosper?
Kathy: What I find is there are a lot of different meanings of the missing middle. What I thought we’d do is talk a lot about the missing middle from my perspective.
As we think about artificial intelligence, it may add trillions of dollars to the economy in the coming years. AI moves the needle on where and how humans interact with technology. Humans help machines, machines will help humans in the future.
In the government, because we have those dimensional challenges I alluded to earlier, when it comes to adopting artificial intelligence, machine learning, and in general, modern technology, we have the mix of aging and new technology. We have aging processes. We have very complex problems to solve. We have a wide range of customer experiences that we have to work with. We have vast amounts of data in various media, and the data is stored in so many different ways.
The public experience with the government is also different because you have all of these different solutions—walking into a local office, completion of forms electronically, manually, mailing forms. We’ve got people calling into call centers. We’ve got research on the website, and so on. This all leads to a workforce […] a variety of workloads.
When I think about the middle, I know that’s a long explanation to get there, but my perspective is how did machines or automation enhance the workforce by shifting us from low-value work to high-value work or by creating new capabilities in an agency?
For example, the modernization of agency websites is enhancing the customer experience. Thinking in terms of our earlier example with usa.gov, the GSA team also enhanced the website in the website in the early days of the pandemic so the public could receive as much current information as possible.
To give some idea of volume from the previous year, they receive, in the neighborhood, 20 million hits on the website. During the first couple of months of the pandemic, they saw triple volumes. The machines are helping us process more requests in this example than we could ever have accomplished manually.
When this work has changed, it also changed how the humans and the workforce are managing information on behalf of the public and how the public is accessing the information. This is a really important perspective when we talk about the middle, how the workforce is starting to adapt to automation, and the automation is adapting to the tasks that we have at hand.
We see these types of changes across the federal government and the result for me is very heartening. The public has a more positive customer experience and we are leveraging new technologies. This in turn creates momentum for more technology, more automation, and more modernization. Think of it as a cog in a wheel or many cogs in many wheels.
For us at the TTS and the CoE, our mission is to design and deliver a digital government with and for the American public. We do this by supporting the agencies as they go through this adoption and the modernization of the technology including creating strategies and working with industries to bring the strategies and roadmaps to execution.
Elizabeth: That momentum and modernization are exactly what we’re trying to accomplish here at FortressIQ. The question is if we drive more adoption, the issue of ethics and responsible AI comes up. How do you ensure that programs have the right checks and balances in place, and what’s the best framework for addressing it?
Kathy: Sure. Thanks, Elizabeth. It’s always a great question. We talked about responsible AI and the ethics of it. It’s always at our forefront. We’re talking about transparency, trustworthiness, accountability, fairness. John talked about trusting the government early in the podcast. The government is continuously looking at how we can ensure the AI meets or exceeds expectations of being accountable and trustworthy.
We need to make sure also that biases are not introduced unintentionally. It goes back to all the different constituencies and the underserved communities that we must support.
Just to give you an example, some of the publicly available databases originating in the ’60s still have biases in tagging women, for example. Women are tagged as housemakers, homemakers, and spouses when they have professions such as scientists, teachers, and accountants.
We were doing our analytics that, of course, skews the information. Responsible AI is also not a one-size-fits-all approach. If we go back to thinking about our RPA solutions, those are very linear in their decision-making. What I mean by linear, they’re ones and zeros, yes and no questions oftentimes.
For example, if you’re using RPAs or bots, as we call them, to route calls to make decisions, the ethics and the bias associated with that are going to be pretty minimal. Think about calling the bank and you press one for account balances, two to pay a bill, or three to talk to a customer service representative. Those are all very simple choices and decisions that the machine is making on there.
If we’re going into solutions that are much more complex, we’re dealing with farm data, weather data, financial information, then the risk and impact to individuals or the risk, in general, is much more impactful. When we talk about responsible AI, thinking about the automation, the type of automation, how the machines are making the decisions, our approach at the centers of excellence is to look at governments, becoming a critical aspect in the implementation and ongoing support of the technology.
Governance, in its purest form, represents the structures and processes that are designed to ensure accountability, transparency, trustworthiness, and fairness. We have a community at practice that helps support agencies looking for best practices on this so that we are applying the proper amount of trustworthiness, accountability, transparency, and fairness oversight to the systems that we’re implementing. In the end, the intent here is that technology remains true to the desired outcome it was originally designed for.
Elizabeth: We ask all our guests on our podcast and would love to hear your unique perspective as well. What advice can you give for the next generation of women in AI and tech leaders?
Kathy: I love this question, Elizabeth. This is such a core question for me. I want you to be passionate about what you do. Be technically curious. I chose to work for the Technology Transformation and Modernization Organization because I believe in technology. I’m passionate about technology. I know that technology can make an impact on the business and constituencies it serves.
This is where we can make a difference to the public that we support. For the young women getting into technology, recognize the role that you play on the teams and be a participant. Be a life-learner, learn from everyone around you, test the limits, and raise your hand.
John: Be technically curious. That’s a great insight and a great point to end on. To recap today’s conversation with Kathy McNeill, the Director of Artificial Intelligence at the GSA’s Technology Transformations Center of Excellence, a framework for AI that leads to successful outcomes includes proper governance, thoughtfully conceived processes based on input from affected stakeholders, and transparency about AI’s role on decision-making. Responsible AI represents this comprehensive approach.
To achieve it, governments must empower leadership and use AI to enhance human decision-making, not replace it. A responsible AI approach should include regular reviews, integration with standard tools and data models, and a plan for potential lapses.
This episode has been part of our special series on Women in AI. A big thanks to Elizabeth Mitelman for spearheading the series and joining the session today. That’s a wrap on today’s show. Thank you, Kathy, for joining us and FortressIQ for sponsoring. I’m John Knisley, and this has been hello, Human.