Executive Interview: Steven Babitch, Head of AI Portfolio, GSA’s TTS
https://ift.tt/2ZonpQq
Primary Mission is to Accelerate AI Investment, Help Agencies Achieve Goals
Steven Babitch is Head of the Artificial Intelligence Portfolio for the GSA’s Technology Transformation Services (TTS), where he is charged to help the US federal government use AI to achieve its mission. He describes four areas of focus for that effort. He brings public policy and private industry perspectives to the task, as a former White House Presidential Innovation Fellow, and as the head of Babitch Design Group. He recently took some time to talk to AI Trends Editor John P. Desmond about his work.
AI Trends: Thank you for being with us today. You were named to an important post in the fall of 2019 as the head of AI portfolio for GSA’s TTS. That came about after an executive order from President Trump in February 2019 that was focused on maintaining American leadership in AI. How’s that going?
Steven Babitch: It’s gone really well. It’s certainly no small feat that we’re undertaking, but we’re making good progress. We’re building out a great team; we’ve got a number of efforts underway. Ultimately, our approach is to make sure that whatever we do, if that’s AI or getting ready for AI or some other way in which the GSA’s Technology Transformation Services can help, we want to make sure it’s aligned with the challenges and priorities of federal agencies that result in them achieving their mission. That’s why we exist in TTS and why the AI portfolio exists as well.
How would you describe the primary mission, and how do you execute that?
Our primary mission within TTS AI is really to accelerate the investment in and the use of AI to help federal agencies achieve their mission. We want to help agencies deliver greater efficiencies, generate better insights, and make better decisions.
We have four areas of focus right now. This continues to evolve as we grow and expand our mission and what we will do for federal agencies.
The first one is implementation and delivery. We’re working with federal agencies directly, such as with the Joint Artificial Intelligence Center within the DoD, as well as with the Department of Labor. We are helping these agencies stand up their AI capabilities and do actual project work.
The second area is the AI Community of Practice. We have a community that aims to help federal employees who are active in or interested in understanding AI and applying it in their context, whether that’s around policy, the technology itself, or building programs or products that accelerate the adoption of AI across government. It’s a platform where agencies can get together, through which we host talks, panels, and workshops, that highlight the great work that’s being done within and beyond government.
We are hearing from the community that they want to know what lessons have been learned by federal agencies working on AI projects. They want to share the best practices from agencies doing similar work. Within the community of practice, we are also forming working groups targeted to specific issues related to AI, whether it’s aggregating resources, providing perspectives or building out tools or other resources. This effort is still early in its development. We want to help agencies on their roadmap to full-scale adoption of AI.
The third area is product development. What we mean by that is, what are the products that can help accelerate the investment in the readiness of AI across federal agencies? One example is a use case library. We hear consistently from agencies that they want to know what others are doing and how they can learn from the experiences. If we can build out a digital use case library, it’s a place where agencies can go to look at the examples and connect with the people who have done the actual work across the federal government.
Another example is an upcoming guide that will help agencies think about and understand what it means to invest in and apply AI. We are looking to build content in this guide that has practical information to help agencies define a roadmap and how to get there, including what building blocks are needed.
The fourth is on external engagement. We want to engage beyond the government to academia, industry, think tanks and others that are thinking deeply about the various issues, challenges, and opportunities for AI. We want to make sure that we fold those external perspectives into the government as well. So those are the four areas that we’re focused on right now.
What would you say is the maturity level of AI in the federal government today?
The maturity level varies across the federal government. We have agencies that are more R&D-driven that have been doing AI research and applying it for years or decades. Then we have a number of other agencies that are less mature. Maturity will vary within an agency. You’re going to have pockets that are exploring it, experimenting with piloting, and moving projects from pilot to scale.
It varies even within those agencies, but there are some that are still relatively early in their investment and use of AI. It’s a mix. We aim to help the agencies move along the path to widespread adoption.
You are coordinating the use of multiple AI technologies across the agencies, including machine learning, robotic process automation, natural language processing. Are there any example related projects that you can talk about?
There’re a couple I could talk about. Within the AI Center of Excellence, we really roll up our sleeves and deliver on projects to help agencies extend their capabilities. For example, we started working with the Joint Artificial Intelligence Center of the DoD in September 2019, focusing on accelerating the DoD enterprise-wide adoption of AI.
One of our focal points is around helping the JAIC set up a premier acquisition office. The government will be doing an awful lot of buying of artificial intelligence; the acquisition office will help to do it in the most effective way. Another focus area is setting up a development security operations (DevSecOps) environment, that will lay the foundation for doing AI across the DoD.
Another project we talk about is within the Department of Labor, which was announced in February. This work is focused on modernizing their agency acquisition capabilities, but this time using robotic process automation. The idea is to start to use RPA to modernize that process, but then expand that over time and scale that throughout the entire Department of Labor as a shared service. Those are a couple of examples.
Is there any company that is a key partner you could identify, such as in the area of RPA?
I think there’s not any one partner per se. I think what’s important for us is that we make sure to certainly stay engaged with industry, and try to fold in and learn from their perspectives and how they implement AI. The US government will be doing a lot of acquisition. We want to make sure we are trying to solve the right problems, then use the right partners to do so.
What would you say are the challenges in your job?
One challenge is to make sure that we’re focused on the true areas of need with the agencies, and then putting a focused effort on getting them to achieve the objective. A big part of that is problem framing, making sure we’re defining the right problems to solve. AI is a set of technologies and tools that help agencies solve problems, but AI may not be the right answer every time.
We are also trying to make sure that, if an agency is not quite ready for the full deployment of AI, how do we get them ready? And if that’s more on the data readiness front, we can help them there. If that’s more on preparing the federal workforce and building up their knowledge and skills and talents and those capabilities, we can help address that as well.
So once we understand the needs of an agency, we can set off in the best direction to move forward.
How do you help the workforce if they need more education or training in how to use the AI?
We’re thinking through that right now. We’re seeing across the federal government right now pockets of training and development of talent. And there’s a couple of key facets to that. One is if we’re going to be doing a lot of acquisition of AI, we need to make sure that folks engaged in the acquisition process really understand enough about AI to evaluate what to buy.
That requires bringing acquisition and technology folks and the mission owners together, literally sitting around the same table to make sure we understand the objective we are all collectively trying to achieve. Part of that will be building out training across those different groups. I would argue that people are the most important asset to focus on.
We’re thinking through right now what are the ways we can partner with agencies to start to pilot and build out the training or the curriculum that can help federal employees at large.
An important piece is, how do we engage with different partners who can bring their different expertise to bear? We’re certainly interested in and open to that kind of partnership. We have some of that within the government, but we could use a portfolio of people who can bring the right level of training and curriculum to help the federal government.
My understanding is that your AI Community of Practice wants to share best practices and tools, the lessons learned, the success stories with the interested professionals. How do you do that? How do you execute on that?
We have been running this for six or seven months now. We have a regular schedule of events, whether they are workshops, or given the time we’re in now with the pandemic, digital and webinar-based events. The effort is to bring the community together on a range of different topics based on the interest level across the federal government.
We are also trying to build out smaller working groups to be more action-oriented, rolling up their sleeves and tackling a particular issue. We are framing that right now, but we are seeing that our team can help galvanize and engage the community to be more active.
Great, thank you. During your time as a Presidential Innovation Fellow, you worked to help the private sector companies conduct user research, translate that into product requirements, test intelligence products to validate value measures of success with metrics. What would you describe as the role of the private sector in relation to what you’re doing now?
Industry is certainly a key stakeholder that we want to and need to engage with the government. We’ll be engaged in a range of acquisitions and procurement of services. So we want to learn more about the tools and technologies that they provide.
In addition to that, we can perhaps learn from industry leaders out there on new ways of building out AI. We need to fold in those external perspectives, in addition to those in academia, non-profits and think tanks. So we see industry as a key stakeholder.
Is there anything you would like to add or to emphasize?
Yes, I would say to those in federal agencies who have a specific problem to solve within a specific time frame, and you are grappling with how to build out the AI capabilities, GSA’s Technology Transformation Services has a range of talented people and approaches that can help you achieve your goals. That is why we exist, to help modernize the federal government and produce better user-centered and citizen-facing services.
We certainly do AI, and we have a range of other tools in our toolbox. We need to make sure we’re defining the right problem to solve, that we’re absolutely focused on helping federal agencies move along their journey to a more full-scale adoption of AI.
Learn more at GSA Launches AI Community of Practice.
Comments
Post a Comment