Adapting to AI: American Competitiveness and the Future WorkForce
Michael Kratsios is the current Chief Technology Officer of the United States. Prior to that, he worked as a Principal at Thiel Capital, Chief Financial Officer at Clarium Capital, and an analyst at Barclays Capital. He completed his undergraduate education in Politics at Princeton University, where he had been involved as the Editor-in-Chief and President of Business Today.
BT (Amy Wang): Based on your undergraduate experience with Business Today at Princeton, how do you leverage your background to execute the role of Chief Technology Officer of the United States?
Michael Kratsios: One of the most important parts of Business Today was understanding how to work with and leverage a team, and that is a big part of what we do here at the White House. We cover a very diverse set of topics, everything from artificial intelligence to quantum computing, and each of those portfolios is very unique. What is critical for us is to bring all those people together, align them to one mission, and deliver policy results for the American people. The efforts that I was part of at Business Today and Princeton taught me the importance of being able to work together, play off each other’s strength, and ultimately achieve outcomes. Princeton has also instilled in me the importance of public service and giving back to our great country. When this opportunity [to serve as Chief Technology Officer] came my way a few years ago, I could not have been more excited to move to Washington.
BT: Since graduating from Princeton, you have also had professional experience working in finance and venture capital. How do those experiences contribute to the perspective that you bring to U.S. technology strategy?
MK: My experience in Silicon Valley opened my eyes to the importance of government regulation in driving innovation in the United States and [how] with the wrong rules, innovation can be stifled ... To me, the opportunity presented in Washington was to be a part of a team and an organization which is able to shape the regulatory posture of the United States to one that is pro-innovation. We are relentlessly pursuing a mission of removing barriers to innovation in this country, whether it's allowing drones to fly, autonomous vehicles to drive, or AI-powered medical diagnostics to be approved by the FDA. We want to create a regulatory environment where innovators are highly incentivized to develop the next great technology here in the United States.
On AI
BT: Under your leadership, the United States’ released a proposed budget for 2021 that includes significant funding for AI projects, as well as a set of principles for federal agencies to meet when drafting AI legislation. Can you describe the need for further advancement in this area? What are the largest challenges the U.S. faces in developing AI?
MK: Our mission since day one has been to ensure American leadership in emerging technologies, and one of the core technologies we emphasized was artificial intelligence. Our initiative was launched by President Trump last year when he signed an executive order launching the American AI Initiative … and it has a four-pronged approach. One is research and development (R&D) leadership. As you mentioned, we have doubled the investment in AI R&D by the federal government, and we made big efforts to coordinate R&D across all of our agencies from the Defense Advanced Research Projects Agency (DARPA) and the National Science Foundation, to the Department of Energy and the National Institutes of Health. The second pillar is around AI regulatory posture. We want to make sure that the regulatory regime we have in the United States is one that encourages AI innovation, while also addressing the important concerns relating to trustworthy AI ... We are the first country in the world to put out regulatory guidance on how agencies should be regulating AI-powered technology. The third pillar is around the workforce. We want to make sure that Americans are prepared to harness the skills of AI in the workforce and to be able to make the most of it as they progress through their careers. So we have lots of programs around reskilling and retooling Americans to be able to take advantage of this technology and also to help prepare and educate the next great American researchers. The last pillar is around international engagement. We believe it’s critical for the United States to partner with like-minded countries and allies to ensure that AI is developed in a way that is reflective of Western values. It's something that we’ve done through initiatives at places like the OECD in Paris and also through my efforts at the G7 and at the G20.
BT: On the topic of AI ethical standards, among the set of 10 principles that federal agencies must meet when drafting AI legislation, fairness and non-discrimination stood out to me. Could you elaborate more on that principle and what motivated that focus?
MK: We want all Americans of all backgrounds to benefit from AI innovation and trust that the technology is fair, unbiased, safe, and secure. We see that researchers around the country are working to develop technical standards for AI fairness, explainability, and transparency. Our administration issued a strategy for how the federal government would play a role in advancing research in these domains, and we deeply believe that increased R&D is going to be critical to making the necessary technical breakthroughs to allow AI to be explainable, to be transparent, and to be robust.
BT: On that point, how do we measure fairness in AI and how do our standards adapt to quickly-emerging technology?
MK: Fairness is very tricky, and it’s something that continues to be researched by our academics around the country. Another way to look at it is, how can you increase the explainability of particular AI models? Once you are able to achieve explainability, you have better insight into the underlying decisions that were made by that particular algorithm. A lot of efforts are being taken to achieve AI explainability. It is something that’s not actually feasible today, and that’s why we’ve invested in programs at places like DARPA to try to make these breakthroughs.
On the Future
BT: You mentioned that the third principle of the United States’ approach to AI is looking at the US workforce, specifically reskilling and retooling Americans to be able to adapt to AI. Under your leadership, there have been initiatives on furthering STEM education. Can you describe how students from both technical and non-technical backgrounds can be prepared to face new challenges and opportunities in technology?
MK: Being able to prepare the technical workforce is one that spans the gamut of all types of education. To begin, you have to look at the K-12 domain, and the president signed a presidential memorandum in 2017 that committed at least 200 million dollars a year to STEM education, with a focus on computer science. And this was actually matched the next day by another 300 million dollar commitment by the private sector. So we continue to focus on getting kids committed early to STEM and to have them be able to progress through their careers with that knowledge as their foundation. The second thing we think a lot about is the makeup of the American workforce. Currently 65% of Americans do not have college degrees, yet we have a federal education system which is primarily geared towards catering towards the 35% who do and what we have emphasized over the past 3 years has been focused on reskilling and retraining all American workers who can benefit from this technology. Being able to design programs that cater to all Americans is a way we can make the biggest impact.
BT: What forms do you think that kind of reskilling would take?
MK: There are a couple; I think one is industry-led apprenticeship and credentialing programs. One thing we’ve identified is that the organizations that know most about the technical needs of the workforce are actually the private sector companies themselves. Bureaucrats sitting in Washington at a desk are probably the least equipped to be able to make the best types of predictions of the skills that the American workforce needs. That’s why we’re working very closely, hand-in-glove, with the private sector to be able to identify the future needs of the workforce and to be able to collaboratively train the existing workforce ... The United States is blessed that it has this innovation ecosystem [that] is underpinned by a set of values … It's one that’s driven by a sort of a free market approach [to] innovation where the private sector is working hand-in-hand with academia and the federal government to drive the best discoveries possible. This is markedly different from many countries around the world and some of our adversaries who have a much more centralized approach to innovation. We continue to believe that this free market approach is one that is ultimately going to ensure American leadership.
BT: Speaking as someone who is majoring in the Woodrow Wilson School of International and Public Affairs, I am interested in technology, though my academic interests might be considered non-technical in a lot of ways. How can students from non-technical backgrounds engage with the changes that are happening?
MK: You need both technical experts and policy experts in order to reach the best conclusion. That's the hallmark of the team we have in the White House. We have the smartest scientists and technologists from top universities in the country working here, but those are married with more traditional policy advisors who can help bring the entire conversation together. When you think about developing AI regulatory guidance, some of it is certainly focused around technical issues, but a lot of it [is] around fair, trustworthy, robust, reliable AI. Thinking through those types of issues and implications at various agencies are ones where technical analysts alone cannot provide the best solution. We think it's a team effort, and we need folks of all different backgrounds to drive good tech policy.
BT: What advice do you have for any students interested in engaging with these technology policy challenges?
MK: I would encourage them to participate in the conversation and be part of the solution. Technology policy is something that impacts everyone’s life. Technology like artificial intelligence is impacting everyone from farmers in Iowa to resource extractors in Texas to pharmaceutical developers in Boston, and as the future continues to progress, there will be few industries that aren’t touched or transformed by technology. I encourage everyone to read up on these issues, to think about and understand the true power of technology to fundamentally benefit the American people but also understand the important ways it should be implemented and that the government can be a part of ensuring that the innovation happens here in the U.S.