 I'm Michael Barr, I'm the Joan and Sanford Wildean of the Gerald R. Ford School of Public Policy. I'm really delighted to welcome you all here for our City Foundation lecture and policy talks at the Ford School event today featuring Tom Khalil. In 2000, the City Foundation, through a generous endowed gift, established the City Foundation lecture series. The lecture series honoring President Ford's long affiliation with City Group brings prominent policy makers from the national and international arenas to the Ford School each year to engage students and faculty in dialogue and to give a public address. Tom Khalil is one of our country's leading experts on technology and innovation policy and it's an honor to have him here with us today. Tom served in the Clinton Administration as the Deputy Assistant to the President for Technology and Economic Policy while doing double duty as the Deputy Director of the White House National Economic Council working on technology and communications issues and nanotechnology initiatives. He continued his role at the forefront of science technology on President Obama as the Deputy Director for Policy for the White House Office of Science and Technology Policy working on initiatives across the range of the White House activities from learning technology, data science, robotics to STEM education. In that role and in these roles and in his other parts of his career, Tom serves as a role model for an advocate of scientific advancement. Let me take a moment to mention many of you in this room are part of the Ford School of Science and Technology and Public Policy program. This is a program that really takes on the ability to help provide our students and our faculty with the tools to analyze complex science and technology policy issues deeply connected to Tom's career and his, I think, distinguished career that I've just given you little highlights of today at his request suggests the advancement, the ways in which science policy really needs to be brought into this broader realm of social science, ethics, values, public policy, and economics. Today, Tom is continuing at work in the private sphere as the Chief Innovation Officer at Schmidt Futures an exciting new philanthropy that Tom will undoubtedly mention. Let me just say a word about format. We're going to have some time toward the end from questions for the audience or as Tom said, maybe closer to the beginning for questions from the audience. I'm so happy to engage with that. We have Joy Rody here who's the Ford School Associate Professor and Interim Director of the Science, Technology, and Public Policy program. Two Ford School students, Celine Seja and Carl Hetchie, who will sift through your question cards and pose them to our guest. For those of you who are watching online, please tweet your questions using the hashtag policy talks. Now it is my great pleasure to welcome Tom Coleal to the podium. Tom, I turn it over to you. Thank you. Is everyone fired up and ready to go? All right. So it's great to be here in Arbor. And Dean Barr, thank you very much for that kind introduction. My father would have been proud and my mother would have believed it. So I'm going to talk about my experience in the working for the Obama Administration in the area of science and technology policy and give you some examples of both specific initiatives that President Obama was involved in launching and reflect on some general lessons learned from that experience. And then I want to make sure that we have plenty of time for questions. So just to say a little bit about how I wound up as an advisor to President Obama and President Clinton, it was really all a path-dependent consequence of a decision I made in 1987 to volunteer on the Dukakis for President campaign. So I wound up working in the issues department, which is sort of like boot camp for people who want to work on public policy because what would routinely happen is that you're given some very short period of time from a few hours to, if you're lucky, a couple of days to try to get up to speed rapidly on some public policy issue and then figure out what is the one to two pages that you think the candidate really needs in order to be able to understand the issue. And believe it or not, this was pre-Google. So dinosaurs were still roaming the earth, so we actually had to call people on the phone. And if you're in the issues department of a campaign, it means that you're responsible for coming up with new ideas so that the candidate can say, if elected, I will do X, Y, and Z. We had to come up with the X, Y, and Z to manage networks of outside advisors so that if some question came up that there was someone that we'd be able to turn to to get the candidate ready for debates to ensure that they are familiar with the issues that voters are going to be asking them as they travel around the country. So in 1988, the candidates would get questions about the auto industry if they went to Michigan or the semiconductor industry, if they went to Northern California. Trying to ensure that the campaign ads were at least reasonably accurate and also we would get flooded with questionnaires by every interest group under the sun, so we would have to figure out which of them we were actually going to answer and which we were just going to punt on. So what you learn how to do is not become an expert in anything, but what you learn how to do is just rapidly get up to speed so that you're at least conversant on the issue in the amount of time that you have. So I did that in 1988. In 1992, I went down to Little Rock and wrote a number of Bill Clinton's position papers and then got a job working for Bill Clinton on the National Economic Council and got a chance to work with Michael Barve when he was also at the NEC in the 90s as we refer to them those dark days of peace and prosperity. So when I worked for the Office of Science and Technology Policy, the main thing that I did was to build a team. So this is the folks that I recruited to the White House Office of Science and Technology Policy. Right next to me was my boss, John Holdren, who is the President's Science Advisor. And one of the people that I recruited is University of Michigan's very owned Shreeder Kota who helped to lead a number of our initiatives in the area of advanced manufacturing. So I don't expect you to be able to see all this, but this gives you a sense for the structure of the White House Office of Science and Technology Policy. We had five major divisions. So one that I led focused on tech and innovation. We had a U.S. CTO that was focused on things like digital services being developed by the United States government, a group working on national security and international affairs. So the link between S&T and national security issues, things like cyber security, and then a group focused on energy environment and overall the role of OSTP was to work on both science and technology for policy and policy for science and technology. So let me say a little bit about what the distinction between those two is. So if the President and his or her senior advisors are making a decision that has a scientific and technical component, our job was to make sure that he was getting the best possible advice. So when Fukushima occurred, he would want to know what are the implications of this, not only for the people in Japan, but what are the implications of this for the United States? Or if there was a year when the flu vaccine was singly ineffective, and so the question he asked was, how do we ensure that we do better in years to come? So that's sort of science and technology for policy. How do we make sure that the President has the best possible advice? And then policy for science and technology, which is what I'm going to spend the majority of my time talking about, things like how much should the federal government be investing in research and development? How do we encourage more young boys and girls to excel in STEM? How do we create an environment that fosters innovation and promotes not only the formation of companies, but the rapid growth of those companies in the United States? So this is the President giving his inaugural address. And one of the things that he said that we were, of course, very happy with is the notion that we're going to restore science to its rightful place. And in part, that was because there was a certain level of friction between the Bush administration and the scientific community, and that arose from a couple reasons. One was this notion of scientific integrity. And in particular, when the scientists would say things like about climate change that were not consistent with the administration's policy, there were certainly instances in which they were being discouraged from talking to the press or that people in the White House were changing the conclusions of scientific reports before they were being issued to the public. And the concern on the part of the scientific community was not, well, on any public policy issue, just listen to the scientists and do whatever they say, because public policy involves juggling lots of considerations, but you should at least make sure that the consensus or the views of the scientific community are being fairly and accurately represented as part of the policy process. So this was something that the President very soon on taking office elevated the role of the Office of Science and Technology Policy and signed an executive order on scientific integrity. So I'm going to talk primarily about some of the elements of a framework for the administration's policies, particularly in tech and innovation, that fall under the rubric of the President's Strategy for American Innovation, which is a document that he released in 2009 and then updated in both 2011 and 2012. And there are three broad areas where he thought that there was an important role for the government. One was for the government to invest in the building blocks of long-term economic growth and job creation, particularly around human capital scientific research and 21st century infrastructure. The second was to create the environment for private sector innovation, because after all it's companies that produce commercial products and services. And the third is the role that innovation can play in helping to achieve a broad range of national goals, like allowing Americans to lead longer, healthier lives or accelerating the transition to a low-carbon economy. So this last part of catalyzing breakthroughs for national parties attempts to answer the question, innovation for what? So what are the major challenges that the United States face, that the world faces, where we think science and technology can make an important role? And so what I'm going to do for the remainder of my talk is to give you some examples, concrete examples of things that we launched to advance this initiative. So one of the things that the President said is if you win the NCAA or the Super Bowl, you get to come to the White House. The same thing should be true if you win a science fair or robotics competition. This was one of the President's favorite events of the year. He had an opportunity to meet with amazing students every year. I remember one year he met with a 16-year-old girl who was already doing research on functionalized gold nanoparticles to attack tumors while leaving healthy cells untouched. And it made the rest of us feel like slackers because at 16, I was playing Dungeons and Dragons rather than working on a cure for cancer. And here he is with Joey, who has developed this marshmallow cannon. And the Secret Service has explicitly told him not to fire it. But the President comes up and he says, hey, ask Joey, well, does this work? So the kid is torn because the President of the United States is asking him to fire this thing much to the delight of the Secret Service. So this was one of the mechanisms that the President used to highlight the importance of STEM education and to inspire more young people to excel in STEM. Towards the end of the administration, he launched an initiative called Computer Science for All, which is making computer science and computational thinking a new basic at the K-12 level. One of the challenges in this area is that in other countries, there is someone called a Minister of Education who can just do this. So in Japan, the Minister of Education can say, computer science is going to become a new basic in the K-12 curriculum. In the United States, we have 15,000 school districts. So in order to make progress in this, there was no one we could call. We literally had to build a movement around this that involved governors, mayors, high-tech companies, schools of education, nonprofit organizations in order to make progress on this. Another thing that we worked on with the Deans of Engineering was something called the Grand Challenge Scholars Program. So this is a program in which undergraduate engineering students can organize their coursework, research, service learning, international activities and entrepreneurial activities around one of the 14 Grand Challenges identified by the National Academy of Engineering. We also, obviously, improving STEM education requires increasing the number and quality of STEM teachers. So we had an effort called 100K and 10, which was an effort to prepare 100,000 high-quality math and science teachers over the next 10 years, and this is a goal that we're on track to meet. And then we launched a series of research initiatives. And one area that a lot of campuses, including the University of Michigan, are moving in the direction of is data science. And so this is an effort that we launched in 2012, really looking at how we could go from data to knowledge to action. How do we go from having, you know, huge amounts of data to deriving insights from that data and then taking some action based on that. In April 2013, President Obama launched something called the Brain Initiative. This was an effort to do for neuroscience what the Human Genome Project did for genetics. So one of the questions that I would ask people in the research community and other stakeholders is, in the same way that President Kennedy decided that we should put astronauts on the moon and have them safely return by the end of the decade, what are the similarly ambitious goals that we should set in the 21st century? And a foundation had pulled together a multidisciplinary group of faculty in neuroscience, nanoscience and synthetic biology. And one of the ideas that grew out of that workshop was, what if we made an investment in tools that would allow us to understand the brain in action. So that would increase our ability to understand how the brain encodes and processes information, and ultimately could lead to new tools that improve our ability to diagnose, treat and prevent diseases of the brain, and to lead to new computational architectures and algorithms that are informed by how the brain works. So, you know, at current directories, supercomputers will require their own dedicated power plant. The human brain only uses 20 watts. So Mother Nature has figured out something really important about low power computation that engineers and computer scientists can learn a lot from. And if you look at the advances that have been made in the area of machine learning, many of these work by training an algorithm, by providing it with sometimes literally millions of examples. And obviously, that's not how a toddler learns, but, you know, a toddler burns their finger on the stove. You know, they don't have to repeat that a million times. And so obviously there's a lot more that we can learn from how the human brain works that could inform the next generation of computer architectures and algorithms. So this enjoyed really strong bipartisan support, and Congress passed a law as part of the 21st Century Cures Act that for the NIH component of this initiative, it also involves NSF and DARPA and IRPA, but for the NIH component provides 10 years of funding, which is very unusual because usually they only provide one year of support at a time. This is the National Robotics Initiative that President Obama announced in 2011, in particular looking at opportunities for human-robot interaction. So what can teams of humans and robots do that neither can do individually? We've got a shout-out for the Materials Genome Initiative from a University of Michigan student here. This was also announced in 2011, and this is aimed at reducing the time required to develop new materials. So it can take as long as 17 to 20 years to go from discovery to high-volume manufacturing of new materials. And I believe materials innovation deserves a lot more attention than it gets. If you think about it, entire epics of human civilization are named after the material system they used from the iron age to the bronze age to the iron age to now we're living in the silicon age. And a lot of the things that we need to do in the area of clean energy, the ability to generate energy, the ability to transmit energy, the ability to use energy are going to require materials innovation. And so this was an effort to figure out how to use new computational tools and informatics and machine learning to reduce the time required to develop new materials by at least 50%. One of the areas that I got really interested in was this idea of incentive prizes. And in the late 90s, I read a book called Longitude, which is about how the British, because the British Navy was losing all these ships, had a 25,000-pound prize back when that was real money in the 1700s. And to develop a solution for a more accurate measurement of longitude. When I worked for President Clinton, I was able to get DARPA Prize authority, which they used for self-driving car competition. The second time they ran this competition, a team from Stanford won. And Larry Page was at the finish line and he promptly acquired the winning team. So this is the origin of Waymo, the alphabet self-driving car effort. So when I came back into government, I was able, working for President Obama, I was able to work with Congress to pass legislation that gives all agencies the ability to support incentive prizes for up to 50 million. And if you go to challenge.gov, you'll see over 800 instances in which agencies have used this. And we referred to this as building on the insight from Bill Joy, who used to say, no matter who you are, most of the smartest people work for someone else. So if you're a government agency, you want lots of people to know what problems you're trying to solve and have an ability to contribute to them. And one of my favorite examples of this is that the Air Force was interested in solving the following problem, which is imagine that you have a vehicle approaching a military checkpoint. And the vehicle is not slowing down like it's supposed to. You would like the vehicle to stop without damaging the vehicle or the occupant. So they put this problem out there and the winning idea came from a retired mechanical engineer from Lima, Peru. And the Air Force only spent $25,000 on this challenge. Now had they used a traditional procurement process, I guarantee you they would have spent a lot more than $25,000. They might not have gotten an answer and they certainly would not have gotten an answer from a mechanical engineer from Lima, Peru. So this is not a substitute for other ways of supporting innovation, but I think it is a really interesting tool and the toolkit that we should have for promoting innovation. We did a fair amount in the area of commercial space and there was a program that was very successful, which was that the United States retired the space shuttle because NASA could no longer certify its safety. At that point NASA had to spend a large amount of money with the Russian government in order to get a ticket for a U.S. astronaut on the Suez rocket to go up to the space station. And there was an understandable interest in having a U.S. alternative. So what NASA did was to partner with companies like SpaceX and they, to their credit, they didn't say, you know, here's the rocket that we want you to build. They said, this is what we want you to do. We want a rocket that will go up to the International Space Station, deliver and retrieve cargo and ultimately astronauts. But exactly how to do that is up to you. And so as a result they wound up for an investment of $400 million getting what would have probably cost $2 to $4 billion using a more traditional approach. So the United States went from being, you know, a laggard in the area of the space launch industry to one of the leaders as a result of partnering with commercial firms such as SpaceX. We were also interested in making it easier for immigrant entrepreneurs who wanted to come to the United States and start businesses to do so. We tried to work with Congress to pass comprehensive immigration reform legislation which would have included something called the startup visa. We're not able to do that, so this is something that we were able to get done through executive action. We worked with Congress on legislation to allow equity-based crowdfunding. So a lot of you are primarily, are probably familiar with donation-based crowdfunding where you might contribute to someone's Kickstarter campaign. This was an effort to extend that to equity-based crowdfunding and also to make it easier for small companies to raise capital and go public without having to comply with all the regulations that may be appropriate for larger companies but may not make sense for small companies. Some of you may recall that the, we had a less than successful launch of healthcare.gov. So what happened after that is that the administration recruited people with technical skills to come to drop everything else that they were doing and work almost around the clock until it was up and running again. And the president appropriately asked, why don't we have these people involved at the beginning of a project? And so we were able to convince a number of the people that had fixed healthcare.gov to stay and they launched something called the U.S. Digital Service which is something that the current administration is continuing to support. A project that Schmidt Futures is supporting is something called Coding it Forward. So this was a project started by students for students. So a number of students were interested in using their summer, students that have skills in computer science and software engineering and design and a number of other areas. And are interested in working in the federal government. And so they launched this project called Coding it Forward which is getting thousands of applications at this point for students who are interested in taking what they've learned in disciplines like CS and design and applying them to important public and societal challenges. Our work was not just limited to the natural sciences in engineering. So we also did some work in the social and behavioral sciences. So the woman in the red dress, Maya Shankar, sent me an email and said that she was interested in working at the White House. And it turned out that she had been a child violin prodigy with Itzhak Pearlman, had won the major Yale undergraduate awards and was a Rhodes Scholar. So I went out on a limb and decided to take a chance on her. And I said, what do you want to do? And she said, well, the British have created this behavioral insights team. I think that the U.S. should have something like that. So in short order, she convinced an agency to house the team. She recruited a team of 20, got the president to sign an executive order, institutionalized in the team, and launched 60 collaborations with federal departments and agencies, taking advantage of insights from fields like behavioral economics that have the potential to improve public policy. And started this work before her 30th birthday. So I want to talk a little bit about some of the different tools. So no pressure. I want to talk a little bit about some of the tools that we had in order to go from an idea to something happening in the world. So one was the preparation of the president's budget. So obviously Congress still has to approve it, but the president's budget is an important starting point for those. So when we had something like the manufacturing innovation institutes that Shreeter worked on or the president's brain initiative that required additional funding, we would ensure that those things that were priorities were included in the president's budget. We would work with Congress on legislation. So I'll already give you the example of the legislation that gave all agencies the authority to support incentive prizes for up to $50 million, or the JOBS Act that created an IPL on-ramp for emerging growth companies. We tried to identify things that agencies could do with the authority that they already had. So the international entrepreneur rule is something that the U.S. citizenship and immigration services had the authority to do, which was to use an authority to admit more immigrant entrepreneurs into the United States. The president also had the ability to convene. So if the president had an event and we invited people generally would rearrange their schedule to come. I remember a lot of times our communications team would not want to have us invite people too early because then worried about what we were doing with leak and they wanted to save the news for the president to announce. So we would routinely invite people on a Friday afternoon over a holiday weekend for an event that was going to occur on Tuesday and without telling them what the event was going to be about and they would still come. And so the president used that not just to get people to come to a meeting but to take some action. So a lot of what we did was to build coalitions to advance particular ideas. And then another thing that we could do is to recruit great people. So Schroeder had a big impact in the area of advanced manufacturing. Maya had a great impact on creating this social and behavioral sciences team. The U.S. Digital Service was all about recruiting the top talent in the industry to come to the federal government and improve our ability to design and build or purchase citizen facing digital services. So these are examples of the tools that we had. And the reason I think this is really important is that policymaking is about creating a coherent relationship between ends and means. So you have some goal that you're trying to achieve and then you're trying to figure out what is the thing that the government can do or that we can challenge other people to do that will help achieve that goal. So we had a whiteboard that was a collection of some of the aphorisms that we had developed over time about how to get things done. And obviously I'm not going to go over all of them but I want to talk about one in particular that was one of my favorite thought experiments which is to imagine that you have a 15-minute meeting with President Obama in the Oval Office and President Obama says, Mr. Barr, if you have an idea for policy issue X, first of all, what is your idea? Why are you excited about that idea? In order to make your idea happen, who would I call and what would I ask them to do? So I'll call anyone in the world. It can be a conference call so there can be more than one person in the line. If it's someone inside the government like the Treasury Secretary, then I can direct them to do something because I'm their boss. If it's someone outside the government, then I can challenge them to do something. So there were a couple reasons for this thought experiment. One is that psychologists have this concept called agency and what it means when you're working in the White House is that you have the ability to send the President of the United States a decision memo and have them check the box that says yes. So that means that there are more things that you see in your environment that you view as potentially changeable because they're the result of human action or inaction as opposed to something fixed like the laws of physics that we really can't do a whole lot about. So how do you give someone else that sense of agency that you as a White House staffer feel? So that's one reason. The second reason is that if you're trying to do something that is complicated, it's highly likely that you're not going to be able to do that by yourself, but you're going to need to build a coalition. So it's not like there is a single individual out there who will be able to accomplish the goal. Well, it's very difficult to build a coalition if you can't articulate, number one, who are the members of the coalition? And again, it could be entities and individuals both inside and outside the government. And what are the mutually reinforcing steps that you want them to take in order to achieve a goal? So if you can say this is my idea and in order to make my idea happen, this is who would need to do what? Then policy makers are more likely to be responsive. I had lots of people who would come into my office and say, we would have some variant of the following conversation. They would say, my issue is important. And I would say, great, what do you want me to do about it? And they would say, you should make it a priority. And so you're far more likely to have an impact on policy if you can articulate this. So answer this question. In order to make your idea happen, who would you call and what would you ask them to do? And then you can begin to assess to what extent the members of that coalition would be willing and able. And if not, is there something you could do to change that? So for those of you who are interested in learning a little bit more about some of the other lessons that I learned, I wrote an article called Policy Entrepreneurship at the White House, which is about how do you have influence without authority that is open access articles. So if you put that into your favorite search engine, you should be able to locate that. And with that, I would be delighted to answer any questions that you have about science and technology policy. Thank you. Hi, Tom. Carl Hesch. I'm a first year MPP student here at the Ford School, and I'm also part of the Science and Technology and Public Policy Certificate Program. So thank you so much for being here. We're gathering our questions, but I'm going to kick it off with this one. What were some of the biggest barriers you faced in accomplishing your goals at the Office of Science and Technology Policy? Well, so one issue is that our founding fathers came up with this idea of checks and balances, which I now increasingly view as a good thing. But you know, when you're in the executive branch, it can be very frustrating if you have an idea and you're unable to convince the Congress that it's a good idea. So one of the ideas that I was excited about and one of my colleagues who eventually became our Deputy Secretary of Education was excited about was to create a DARPA within the Department of Education. So as you know, DARPA was created after Sputnik and is invested in things like the internet and stealth technology. Many other agencies lack a similar capability to support high-risk, high-return research, and we thought that there was a case for doing that in the Department of Education given the importance of education to our well-being and to citizenship in the 21st century and economic growth. But we were unable to convince the key members of Congress that this was a good idea. So that's certainly one thing that certainly I understand the need to have this sort of division of power, but when you're in the executive branch, it can be very frustrating. Thank you for that response, Tom. My name is Selene. I am a first-year MPP with the Ford School and part of the Science, Technology, and Public Policy program. From the audience, given today's political climate, what do you see opportunities for people who want to affect science and technology policy on a larger scale? Yeah, so one of the things about science and technology policy is that there's not necessarily this sort of clear career trajectory that exists in some other areas. I think one route that a number of people have used is the AAAS, S&T Fellowship Program. So that is one mechanism that people have used to sort of get some experience in science and technology policy. And I would say that the administration is taking a fairly hands-off approach with respect to many of the science agencies, like the National Science Foundation and the National Institutes of Health. So there's good work that is continuing to go on in the science agencies. I know that there are some states that are creating fellowship programs at the state level, so I don't know if Michigan is doing that or thinking about doing that, but the Moore Foundation, for example, has been supporting a S&T Policy Fellowship at the state level for both state agencies and the state legislature. Thank you. Now we have an intellectual property question. What are your thoughts on companies that buy patents to sue other companies and how can we reform this system? Well, this is not really an area that I am an expert. I know that intellectual property is difficult because you're trying to balance competing goals. On the one hand, you want to have sufficient protection of intellectual property so that firms are going to be willing to invest in research and development. I think one problem that we have with our patent system is that it's what it's referred to as a unitary system, and I think that leads to some problems. What I mean by that is that intellectual property plays different roles in different industries. For example, in the pharmaceutical industry, I think you can make a much stronger case that in the absence of having this period of exclusivity, a drug company is not going to invest $2 billion in getting a new drug candidate all the way through clinical trials in the absence of exclusivity. In other industries like IT, they are competing much more on the basis of time to market, and IP is used more for defensive purposes rather than anything else. I think it's worth exploring whether or not we should look at the role that IP and patents are playing in different industries and being able to have something that is more tailored to the role that IP plays in different sectors. Earlier today you mentioned two market-based approaches, Space X and NASA and the Inceptive Prize. Are there any specific industries that a market-based approach is uniquely suited to advance versus government? There's a set of tools that the global health community has developed that I think are really interesting and should be used outside the global health community. They have had to deal with the following challenge, which is that there are innovations in global health that have a high social return and a low private return. A canonical example of this is vaccines for diseases of the poor. Left-of-the-own devices drug companies won't work on vaccines for poor people. So an economist by the name of Michael Kramer came up with a clever solution for dealing with this problem, which is known as an advanced market commitment. This is essentially the governments or other philanthropists saying to the drug industry, if you develop a vaccine which is safe and effective, then we will purchase X million doses at Y dollars per dose. And so the reason that finance ministers really like this idea is that had the drug companies not delivered, the governments would have been out zero dollars, right? So the way the federal government is currently set up, the government tends to make financial commitments that are contingent on failure. So we have 2.6 trillion dollars of loan guarantees on the federal books. So that's the government saying if you go bankrupt, then Uncle Sam is good for it. With this advanced market commitment, the government is making a financial commitment that is contingent on success, right? And saying, hey, if you develop a vaccine which is safe and effective, then we'll buy it. So I think we should be doing a lot more of that. But what it requires is the capacity to do 3 things. One is to identify an unmet need. You know, in the case of a vaccine, it's like million kids under the age of 5 die every year from a vaccine preventable disease. No vaccine currently exists. So that's the unmet need. And the third thing is to develop a performance-based specification. So again, the government is not saying, is not dictating the how, it's just describing the what. So in the drug industry, for example, this is called a target product profile. This is a description of what it would mean for there to be a safe and effective vaccine. And then the third thing is if there's a market failure, if there's a large gap between the social return and the private return, then what type of incentive would be necessary to get the private sector to work on this problem? So again, that approach is not going to work in all instances. But particularly in those areas where you can have a reasonably clear definition of the problem and what an effective solution would look like using these approaches like incentive prizes or milestone payments or advanced market commitments may be one way of addressing those. Next question is, in terms of long-term security and stability specific to global competition and national security, how do we weigh risks in the speed and open accessibility of emerging technologies while still attempting to maintain our position as a global leader? Very carefully. I think one thing is that when you are dealing with emerging technologies to really think at the beginning about what the risks could be associated with these, and in some cases that may lead to some selective departures from open publishing. And so an example of this is that there was a big debate in biomedical research associated with something called gain of function. So for example, if you're trying to understand why a virus or a bacteria would be more virulent, on the one hand you want to understand what those mechanisms are so that you can be on the lookout for that, but on the other hand if you just publish that, then there is the risk that that knowledge could be misused. So I do think that there are some instances that the National Academy of Sciences called Experiments of Concern where there are clearly dual-use applications associated with these and that you may need to have a review process either before the research is funded or certainly before the research is published. We just saw an instance of an organization doing this voluntarily, so OpenAI published some results on the use of natural language processing to generate fake news and they said we're not going to publish the algorithm, we're going to publish, you know, we're going to describe the results and we're going to go around and brief policy makers so that they have some sense for how this technology is occurring. Another example of a regime or a set of norms that the research community has come up with is if you identify a vulnerability associated with an operating system or some other form of information technology, let the vendor know, give them an opportunity to fix it before publishing it so that you don't have a situation of zero-day exploits. So these are examples of an effort to balance some of these considerations, but it is a really important area and unfortunately there's no set of principles that you can articulate that will allow you to figure out what the right thing to do is and reasonable people will disagree about where you should draw the line. Number one and number two, we will inevitably be surprised. So we may be able to come up with some examples of how this technology will be both used and misused but I think a certain amount of humility is needed because we might have some theories but we will inevitably be surprised by what happens in the real world. So great question. Thank you. In the last couple of months we've seen growing cybersecurity breaches, Facebook's Cambridge Analytica and the deterioration of data privacy. What is the role of policymakers and the federal government in ensuring that this doesn't happen again, if any, or do we expect Facebook to self-regulate? Yeah, so certainly other countries are not taking that approach so that is not the approach that the European Union is taking. So they have launched comprehensive privacy regulation through something called the GDPR. And another thing that people have looked at is increasing the potential sanctions that are associated with various data breaches. So one way to get companies to do less of something is to increase the penalties that are associated with that and a minimum to ensure that they have a duty to disclose when there has been a data breach. So I think that the pendulum is beginning to shift with more companies actually saying there probably needs to be a national privacy framework. And in part what they're interested in heading off is having each of the 50 states have their own state level privacy legislation that would make things a lot more difficult for them. We've seen as technologies emerge and as big data increases that vulnerable populations can be disproportionately affected. How did your office prioritize or consider these effects and vulnerable populations and how can we do better moving forward? Yeah, so we issued several reports looking at the interaction between big data, machine learning and privacy and civil liberties and vulnerable populations. And let me give you one example that people are concerned about. So as many of you know, the way these recent advances in machine learning have occurred is that algorithms are trained rather than programmed. So if we want an algorithm to be able to distinguish between a cat and a dog, we give the algorithm lots of examples of labeled training data. And it in one case forms this network that is constantly adjusting the weights between the nodes in the network until it does a highly accurate job of mapping between an input and an output. So one thing that people have noticed is that if the training data itself reflects existing biases, then algorithmic decision making can reinforce those biases. So there's now an active research community called Fairness, Accountability and Transparency that is trying to address these issues at the design stage. So the research community is not saying, oh, this is someone else's responsibility. They're saying we should be addressing these issues head on that are generating research results to both understand if these biases exist but also to come up with some mechanisms for addressing them. Many science and technology policy scholars worry that the decentralized nature of science technology policy has resulted in a research system that provides good support for research but doesn't always connect that research to social outcomes. For example, great medical research but lack of access to high tech care. How can we do a better job of innovating for the public good? Great question. So one of my concerns is that the ability of different federal departments and agencies to interact with the research community varies widely across the federal government. So there are five agencies that account for roughly 90% of the federal R&D budget and that's the Department of Defense, NIH, Department of Energy, NASA and the National Science Foundation. So agencies that have the responsibility for worrying about the bottom half of the income distribution, so agencies like HUD or the Human Services Parts of HHS or the Department of Labor have little or no capacity to interact with the research community. And so I think an interesting thought experiment is to imagine that one of those agencies had a research arm and ask, number one, what goals would it set? And number two, what are examples of projects that it would support in order to achieve those goals? So we just put out a working with a nonprofit organization called Jobs for the Future, a call for ideas for what we're calling a unicorn for the middle class. So as you know, in Silicon Valley the status symbol is a startup that has a market cap of a billion dollars. We challenge people to come up with ideas that would increase the incomes of 100,000 non-collegiate workers by $10,000. So that's a unicorn for the middle class. So there are certain types of research questions that are not even being asked because there's no private sector incentive to invest in these. Number one, and number two, the relevant part of the federal government, the relevant mission agency like the Department of Labor, for example, has limited or essentially no capacity to interact with the research community and think about science and technology as one of the potential tools. So for example, if you talk to people about housing policy, you'll get into a discussion about zoning and building additional public housing or subsidies or something like that. The way an engineer would think about that problem is, well, how could you make the house itself less expensive? Now that might not be the right answer, but my view is that it's at least one of the ideas that ought to be considered when we're thinking about solving some of these problems. And because of this imbalance, a lot of these questions are not even being asked. And so I actually think there's a role for research universities to go to their alumni and say, hey, we think there's an opportunity here. The federal government isn't immediately going to fund this, but we have an opportunity to demonstrate what might be possible if we made this societal goal an active area of research. What differences have you found in your ability to shape or impact S&T policy from OSTP versus philanthropy like Schmidt Futures? Yeah, so the federal government has a different scale. So one of the things that I worked on for President Clinton was the launch of something called the National Nanotechnology Initiative, and this has resulted in a cumulative investment of $23 billion in nanoscale science and engineering. So the federal government has an opportunity to operate at a scale that is different from private philanthropy. Now the advantage of private philanthropy is greater flexibility and speed. So someone who is working for a philanthropist does not have to get 60 votes in the United States Senate. So that I think allows a philanthropist to respond more rapidly to emerging challenges and opportunities. But private philanthropy is in no way a substitute for federal investment in science and technology because of the, I think, necessity to have, you know, democratic voice in public priorities and just the different scale that the federal government operates. You know, the executive branch agencies have a total budget of $4 trillion, so that's not something that is going to be replaced by private philanthropy. From our viewers on the live stream and Twitter, how can we filter fact from fiction, fake news, information in this world of mass data and communications? Any advice? Oh, I got an algorithm in the way. I'll email it to you. You know, there are a number of interesting experiments that are going on. So one experiment that I'm aware of, for example, is trying to leverage training in critical thinking and crowdsourcing. So training people in critical thinking and the ability to, you know, tag or identify weaknesses in arguments and just training a lot of people to do that. So I don't think that, you know, there's going to be a quick technical fix to this, but I think it is an active area of research. And, you know, I have to say that in the 90s, I was really excited about this idea that we were going to have, you know, the equivalent of a printing press and a radio station and a TV station for everyone who was, like, connected to the Internet. And, you know, I think I'm now more aware of some of the challenges that that poses. What advice would you give for people not working in scientific fields who want to affect change? So, you know, Dean Barr will tell you that effective public policy requires people coming together with lots of different skills. So, you know, I just talked about this idea of advanced market commitments. That's an idea that came out of the economics profession. The, you know, whole idea of taking insights from behavioral economics and applying them to public policy. The idea of mechanism design, which has been used to reduce the waiting line for the kidney transplant. So lots of different disciplines have the ability to contribute to public policy. The other thing is that you need people who can help translate between scientists and engineers and policymakers and the public. So many times scientists and engineers are not particularly good at explaining the importance of what it is that they're working on and having people who can understand what they're doing and why it's important, but still have the ability to communicate with the public can play a really important role. We talked about the rise of AI and machine learning and we know that it's growing at a very rapid pace. Yes. There is contention within the public and the American system currently that robots are going to replace people in the workforce. Kind of how do you quell those, I guess, specific myths and what role will robots have in the workplace in the coming future? Yeah, so some of the analysis that has been done is an effort to try to address that question at the level of tasks rather than jobs. So to say what are the tasks that people are currently engaged in and which of those tasks are susceptible to automation. And some of the analysis that's been done, I think, is concerning particularly for non-college educated workers. So it's not to say that everyone will not be affected by this at some level, but it would be particularly hard for non-college educated workers. So I think that's an area that deserves a lot more attention. The United States is near the bottom when it comes to public sector investment as a shared GDP in what economists call active labor market policies. So this is things like help with job search and job training and reskilling. So I think that's definitely an area that needs a lot more investment. But we need to figure out how to make that effective. So I wouldn't be in favor of just throwing a bunch more money at the current system. I think we've got to look at the workforce development system and figure out how to make it more effective. Because I think in many cases there are not very strong connections between training providers and employers. So the training providers are running someone through a program, but there's no guarantee that they're going to get a job at the other end. So I think we have to ensure that the workforce development system is a lot more demand-driven than it currently is. To build off that question, how can the U.S. government better serve tech companies through H1Bs or reform or improvements? Yeah, I don't necessarily think it needs to be done through H1Bs. So I think one of the problems with H1B is that it creates an unequal power dynamic between the employer and the employee. So President Obama had a number of proposals to just increase the number of green cards for highly skilled workers. So one of the ways that you can do that is to have every visa that you're providing not only cover the worker, but his or her family. So that's one way of doing that. So I think there are ways to increase. So the problem with the H1B visa program is that it's A, it's temporary, and B, that it creates this unequal power dynamic between employer and employee. So when we looked at this issue, our view was just to increase the number of green cards and also make it easier for graduate students who are getting a STEM degree if they want to stay in the United States, we should staple a green card to their diploma and not force them to go overseas if they want to stay in the United States. During your tenure at OSTP, what was involved in initiatives to ensure the public has free open access to outputs of scientific research? What do you think is the role of government in promoting open access to knowledge and what is the role of foundations like Smith Futures? Sure. So there had been a policy that the NIH and Congress pushed, which basically said, if you're taking federal money for your research, then the publication should be open access within a year. And so we extended that policy to the other major science agencies. So that was one thing that we did. And the second thing that we did to say that this policy should not just apply to the publication but to the underlying data. Now, you know, that's been more difficult. The publications is a lot more straightforward. So agencies are taking steps in this direction, but that's an area where I think we have a lot more work to do. And I think that the case for doing this is even stronger given advances in machine learning because what machine learning in certainly if you're talking about supervised machine learning is driven by access to data. And so if we make that data available and we make it reusable, that's going to accelerate the pace of scientific research. This will be our last question. In working in S&T policy, how important are soft skills, including communication and team building, compared to scientific literacy? Very important. Yeah. So some of the important soft skills, as I said, are the ability to communicate across the technical, non-technical divide. I think that's one. The second is the ability to build coalitions. You know, there are very few things that you do by yourself that requires eliciting collaboration and cooperation with people that do not work for you or do not report to you. So the ability to do that is really important. The ability to develop relationships that are based on trust, mutual understanding and reciprocity. So I would not be effective if I was constantly sending a one-sided stream of requests from me to other people in the federal government. It would be like if you had a friend when you only heard from them when they needed help moving. I would get old after a while. So one of the things I would do is I would have coffee or dinner or drinks with people to try to understand what it is that they were trying to accomplish and how I could help them. So it was not just sort of like me constantly calling up and saying, hey, you know, can you do X, Y and Z? So I think that those skills are very important. And there were, you know, a number of brilliant scientists in the federal government who I think had challenges because they were less strong in some of those areas. And so as a result, we're less effective. And if the piece that I wrote on policy entrepreneurship talks about some of the skills that are particularly helpful. That's really great, Tom. I'm not sure whether this is working, but we can also get, I think, that article around for people to see and they will post on the website as well. Please join me in thanking Tom Taylor.