From service delivery to data accountability to issues of equity, digital technologies are having a tremendous impact in the public arena. David Eaves, lecturer and faculty director of digitalHKS , answers callers’ questions on the resources, tools, and experiences needed to identify the risks, opportunities, and impacts of digital technologies on the public good.

Wiener Conference Calls recognizes Malcolm Wiener's role in proposing and supporting this series as well as the Wiener Center for Social Policy at Harvard Kennedy School.


Mari Megias:                

Good day everyone. I am Mari Megias from Alumni Relations and Resource Development at Harvard Kennedy School, and I’m delighted to welcome you to this Wiener Conference Call. Today we’re very lucky that David Eaves is joining us. David is a lecturer at Harvard Kennedy School and faculty director of the digital HKS initiative, which focuses on digital technology and the public interest. He’s a public policy entrepreneur and expert in information technology in government. As a former advisor to the mayor of Vancouver, David proposed and drafted one of the first open data portals in Canada and the world. He has advised the Canadian government and other governments on open data strategy, and he served as the first director of education for Code for America, a nonpartisan, non\political 501(c)3 organization that addresses the widening gap between the public and private sectors in their effective use of technology and design.

In his work with the U.S. government’s general service administration, David helped various federal agencies improve how they serve the public through technology. In 2018, he was named one of the world’s 100 most influential people in digital government by Apolitical, a free platform that helps public servants partner with each other to solve hard challenges. We are very fortunate that David has come to talk to us in today’s Wiener Conference Call, which I’d like to note is on the record and will be recorded and posted online after the call. David ...

David Eaves: 

Thank you so much. It was great to get a briefing on the questions that people had beforehand, and so I am somewhat tempted to try to limit my remarks so that we can do more on the questions. I want to serve the people who are on the call, and not just talk about the things that I care about. But let me say two or three things that I’m working on here at the Kennedy School to provide some context before we do that. 

I teach several courses here. One of the key things I’m really trying to think a lot about is, how do we increase the state’s capability to use technology in effective ways, particularly in the provision of services? I’m particularly concerned with those who tend to be more marginalized, but really to all citizens.

One other big thing I try to share with my students is drawing some metaphors about how the state has currently been using technology, and where I think that we are going, what the roadmap seems to suggest is the direction that we’re going. The metaphor I often use for this is, the way we currently build technology in the state is that we try to build these giant monolithic systems. This is the equivalent of building like a ... You want to get from Boston to Providence, you build a whole highway to get from Boston to Providence. And on that highway, there are a whole bunch of capabilities and features that are useful to you. But then let’s say you want to go from Boston down to Foxwoods Casino, which is basically on the way, and a little bit farther. We build a whole new highway that will run parallel to the existing highway for 70 percent of the journey, and then divert maybe for the last 30 percent. So we don’t leverage the work that we’ve already done in building this new service. We build an entirely new expensive service. And then we’ll need another one, we want to go to New York, whole new highway all over again. So we have these big monolithic systems that are very, very hard to adapt, very, very hard to alter, enormously expensive to replace, and in fact sometimes so expensive to replace that basically it’s cheaper to just keep deferring the replacement of them almost into the forever future. And part of this matter is because ... Helen Margetts at Oxford has this great line in one of her books about ... She really says that, the options of a politician or a policymaker, whether they’re conscious of it or not, are really limited by the technology capabilities that they have. So if you want to pass some new law that has some sort of enforcement mechanism or has some compliance piece to it, you are inevitably going to have to build some mechanism by which to capture that information and store that information.

And if doing that is a six- to eight-year process, I mean, one of the secretaries of state in California tells me, “The governor can say, ‘Let’s do x,’ and he or she won’t even see that occur until after they’ve left, if they’ve left.” The governor said eight years later, he’s like, “This is just crazy.” And so people look at the technology capabilities they have, and basically say, “I’m going to assume I can’t add to them.” So the policy options that are available to me are constrained by whatever I can try to get the technology that I have to do for me right now. And usually that’s a very limiting set of possibilities. So figuring out how to increase the state’s capability to deploy technology to serve policy and public’s needs, I think, is enormously important.

So where are we going with that? Well if you look around the world at some of the states that I think people would argue are at the cutting edge, these would be states some of you on the call may be aware of. States like Estonia, frankly, and, with some controversy, India as well. What they are trying to do is to replicate some of the work that’s been going on in the private sector, particularly with the new generation of companies that have emerged. So think of your Facebooks and your Googles, which is to really figure out what are the core services that you need to build this technology solution around which then can be leveraged by other services. So we refer to these as platform services. And the idea here being that when you look at a government service, almost all government services require you to authenticate who you are. Am I David Eaves? Do I have an email address and a password to validate? Social security number? I need to collect that social security number. I probably want to validate or know what your address is. I’m going to check that against some sort of address database.

All of these systems that we built, rebuild that functionality over and over and over again. There’s incredible duplication. It’s also why even with, say, something like the IRS, if you have multiple relationships with them, say as an individual, but also as a business, you have different usernames, different passwords. And they ask for the same information, but in different ways. There’s no standardization. These governments are really trying to figure out, how do we build an authentication system that can be deployed into any service that you use> So you create a very, very consistent experience, but more importantly, you only build this once, and then you leverage it over and over and over and over again. So this is a little bit like, while there might be some concerns about this specific example, but you can use your Facebook account to log in to lots of different services, or your Twitter account. 

That’s kind of like a platform identity service. What would it look like for the government to provision those, and only have to build that once as opposed to over and over again? That is directionally where I see a lot of governments going, and that’s exciting because in places like Estonia, where they have a lot of these platform services, things like your identity but also addresses, all sorts of social services now—to add a new service is actually quite cheap, because now, rather than building all these things, you can just go pull the API. But you can go pull the service for identity, the service for this, the service for that, merge those together, and then just build the small component that you need to build that’s the unique information that you need to gather, or the unique thing that you need to do. And this takes what is often in the United States, billion-dollar problems, and turns them into $500,000- to $2-million problems. 

So it allows them to do a lot more experimentation, and to try services out in a way that would be so expensive that you’d never try to do that here. So that is directionally ... I get pretty excited about that world, and I spend a lot of time trying to talk to my students about platforms, or what they look like in government. I think the cautionary piece is ... One of the things I spend a lot of time thinking about is, I have what I call the “what happens if we win” conversation. So let’s imagine a world where now, the United States say, adopts this approach, is actually successful, executes it. What are the new problems that we’re going to have? We’re still obsessed with the existing problems. The danger is always that you don’t think about what are the new problems that this will create, and there are three or four that are really sticking with me as quite problematic.

The first is, I do worry about this world breaking the social contract between the state and the individual. Some of the trust that individuals have in the state is based on what I call “perceived incompetence.” So you might be comfortable sharing information with the IRS because you know the IRS does not have the capability of sharing that with some other organization for whom you don’t think you’re in violation of rules, but you wouldn’t want them poking around and doing a deep audit on you for that. And so people share information because they perceive the government is not very effective at being able to build a profile about you, and understanding a lot about you, and doing predictive analytics about who you are, and what you might want to be doing. I think in a world where the state actually becomes quite effective, platforms would get them quite a long way there, you might suddenly discover that citizens are much, much more reticent to share information with the state, much, much more cautious. And that could actually impede the state’s ability to do some of the functions that we think are important, and could actually erode trust that people have in the state. So that’s the first concern that I have.

The second concern I have is in this world of platforms, it really breaks the siloed nature of our government, where you have these ministries in which this monolithic system exist because they exist within Homeland Security, or they exist within Health and Human Services. And then when you start thinking about platform services, they really span all of those ministries now. You’re deploying identity authentication into every single ministry off a single platform. And so it means whoever controls that system, may be able to exert influence across government in a way that presently is very hard for us to imagine, or exists exclusively in the White House but with limited capabilities. So you could imagine that a government saying, “Oh you’re getting ready to deploy a service that we don’t like, we’re going to deny you access to these platform services so you simply can’t deploy it.” Or, “The budget required to do that would be so bad that you would never get it because everybody would assume that you couldn’t have it. So we won’t say no, we just won’t let you. And then you won’t have the money, and it just will never happen.”

And you can imagine all sorts of political choices being made where certain types of services that are not aligned with a particular ideology now get denied because they have this central lever if these platform services are centrally controlled, or under the control of a particular player. And we shouldn’t assume that government won’t exploit those. The example I love to use with my students is, I always ask them, “What is the drinking age in the United States?” And everybody always knows the answer, which is 21. But if you ask them who regulates drinking ages in the United States, they get more flummoxed. A lot of people are like, “Well it must be the federal government.” But it’s the states. And doesn’t anybody think that it’s weird that the 50 states who just can never agree on anything, just magically aligned around 21 as the drinking age? And it’s because ... I believe it’s under Reagan when they renewed the Highway Act, they put a rider in there that said, you only get highway funding if you make the drinking age 21. 

This was a way of the federal government using its control over a platform, which was the funding of highways, to exert policy control in a domain where it really has no jurisdictional business. So if you started to imagine a set of digital tools, not just budgetary tools, but digital tools that lots of players become dependent on, what power are we giving the people who own those tools. 

The third thing I get worried about is, who owns control of those tools? Not from a governance perspective, but from a technological perspective. If these tools are owned by private actors, then the roadmap that those private actors have will then dictate what the capabilities of government are. There are a number of very powerful companies that already provide technology to governments, so you say like a company like Esri, which provides geospatial information ... actually grew out of the GSD here at Harvard. Huge numbers of cities use them as the platform for storing geospatial information. Whatever they choose to put in their product or to remove form their product has a huge impact on what the capabilities of cities are. 

So influence over that roadmap, or control over the roadmap of what that product can and can’t do, then goes on to have all these downstream effect. So if we become a government that depends on these platforms, then the ownership of those platforms from a technology perspective will also matter dramatically. 

And then the final piece that kind of links to an earlier piece I said, which is not just ... I also worry about what does this mean from a multijurisdictional perspective. In the earlier example, I talked about maybe the federal government asserting control over other ministries. I kind of alluded to the state example, but you can then begin to imagine, well what if everybody’s dependent on the federal government for maintaining a social security database that is a unique centralized one, and that almost all services are linked to it. Does that mean that the federal government cannot exert control over local government? So they need access to that social security database to provide food stamps or other types of services. 

And you know, say there’s a disagreement about sanctuary cities, could the federal government say, “Well we’ll simply shut down your access to the social security database if you don’t decide to end your sanctuary city status,”? That kind of centralized power that can reach right into your operational capability is a very interesting threat model that we would want to think about in this world. So that’s kind of this first piece I wanted to just address about the excitement that I have about this world, but also the real dangers that we need to be thinking about if we actually are successful in executing around this world. 

Maybe the second big thing I spend a lot of time thinking about here, we just organized a conference about it this past weekend, and there was a Forum event on Friday, which you can actually watch the video which had Reid Hoffman, and myself, and Ash Carter, Latanya Sweeney, and Vanita Gupta who’s a civil rights lawyer, is this notion if public interest technology. 

It grows out of a group of people who are starting though think more and more about what is the responsibility of technology as in how they create the world that we’re all going to live in. Here the impetus is a little that comes out the notion of the public interest lawyer that emerged in the ’70s and the ’80s, which was ... The feeling was before that, most people who went to law school went on to work for one of the big law firms. People didn’t really conceive of careers outside of corporate law so much. And then, some students started to open clinics that were trying to serve local populations and really think about human needs in law. And that led to this notion of public interest law, and a set of courses that were like, Hey, can we also think about the law and the environment, the law as a tool of social justice, the law as a tool to try and go and create ... to improve the lives of marginalized communities. And it created several things. It created a whole set of career paths now where lawyers can envision themselves not as purely corporate lawyers. It created a whole set of organizations that use the law to try to pursue their goals. And it changed the notion of what it meant to be a lawyer. Now in many law schools, doing clinic is actually a requirement to graduate. And so there’s a sense of what would the analogy be in the public interest technology world? And here there’s kind of two sides of this coin. I think there’s real interest in thinking about, well we need to go to the computer science departments and think about what does it meant to bring public interest to those places. I’ve drafted how I think we should be teaching technology to policy students, which you can find online in a Medium post. It was done exceptionally well, and our students find very, very helpful. 

But really wrestling with this notion of how do we do both of these things. What are the skills that the technologists need to gain, and what are the skills that policy people need to gain in order to serve more effectively? And I think anybody who watched the Facebook hearings can see the enormous gap, if only that, at the legislative level the sophistication of the questions is just so low. Two years ago the Europeans did this, and two years ago their questions were far more sophisticated than what was done here in the United States. So there’s this huge gap here that we need to find a solution for. And the big challenge I see around this is, and part of my mission here at the School, has just been trying to get people to be conscious of the world that we’re going into. 

My friend John Lilly who works for Greylock who was the CEO of Mozilla, he told me this story that I know tell to all my students. It’s a David Foster Wallace joke that he told at the beginning of a commencement speech where he says, “There are these two fish swimming in the water, and this bigger fish, older fish swims the opposite way. And as it’s swimming by them, the older fish looks at the two smaller, younger fish, and says, ‘Hey boys, how’s the water today?’ And the two small fish keep swimming along, and a couple minutes later one of the small fish turns to the other small fish and goes, what’s water?” And the point here is, people try to think of technology as a tool or as a thing, and it’s not really that, it’s the medium in which we’re in. Even when we’re not using our computers, you live in a world where a lot of the decisions are modulated or even just determined by technology. Even the room we’re sitting in, these lights will auto dim at some point. It will get dark.

This physical medium is controlled by computers. I think there are three groups of people out there. There are the people who know they’re in water, and are really good swimmers, and I think they often are ending up in the private sector, and we need to help them think about what public interest looks like, and what their central responsibility is. There are the people who know they’re living in water, and feel like they’re drowning, and we need to give them better tools to think about what it means to be effective living in water. And then I think frighteningly there are a large number of people who don’t know they are in water, and think they’re on land, or somewhere else, and we need to make them conscious of the fact they are not in the medium they think. They’re in a very different medium. And the actions they take will have very different outcomes that the ones they are thinking

So I spend a lot of time thinking about that. So maybe with that as kind of an introduction, I don’t want to talk too long, I’d love to get to questions make this interesting because I have no idea if people on the call are finding this interesting, so let’s get to their questions and be helpful.

Q: What comparisons can you draw between the industrial revolution and the current digital technology paradigm shift? Thinking specifically of workplace safety, child labor, population movements, intellectual property, and wealth distribution and redistribution, and the legal aspects thereof.

So I think there are a lot of parallels. I think the stronger ... Maybe I’d like to reframe, and say maybe the stronger parallel for me at the moment is less to the industrial revolution and more to the revolution brought by the printing press. And the reason I say that is because while I do think the industrial revolution was about massively increasing our capacity to generate energy, and then as a result of that, our productive capacity to create physical goods. The printing press was much more about a massive increase in our ability to generate and store ideas. In some ways, the printing press was like the hard drive revolution of its time. Where you know, before hard drives, books were really, really expensive, and because they were so expensive, very few people could read them. There was a very specialized technician called a monk who knew how to go and read those books, and write those books. 

As a result of the printing press, the cost of storage dropped precipitously, which suddenly meant all sorts of laypeople could suddenly gain access to that knowledge and record that knowledge and write, and read from those hard drives. And that kind of feels to me almost more analogous to the world that we’re going through. And I think there was some ... I remember writing a piece about this a while ago. I kind of want to revisit it actually. It’s just a lot of people in that metaphor really focus on the good parts of the printing press revolution, which is that the narrative is a small person, think Martin Luther, comes up with idea, is able to print and distribute it, hammer it to the door of the church. People focus on the mobilizing and democratizing notions of knowledge, and that all sorts of people now could read, and it spread wealth much more broadly. 

I’m very excited about those things, and I think the internet has done similar things. I just think the printing press also has a darker side to it, particularly around state power. And I often talk about you know, if you think of a French person living in France a hundred years before the printing press existed, their notion of identity was very, very local. They were notionally French, but France was this thing, and the king was super far away. They probably spoke some weird dialect of French that really other people didn’t speak, if they even spoke French. And their relationships were all like their local lore. And then if you look 100 to 150 years after the printing press, Napoleon is able to martial a million people, and march them into Russia. The reason Napoleon can do that is because they standardized language, they standardized history. They erase all sorts of people, they standardize what it means to be French, and really the printing press enabled the creation of nationalism, which becomes a tool that the state was able to use to extract massive amounts of social capital from its population, such as mobilizing a million people into an army, which was impossible to imagine 200 years earlier.

Other things happened there, but those things were pretty crucial. And so when I look at what’s going on on the internet, a lot of people talk about the state being slow, and not being ... I’m like, “The state’s always slow, until it’s not slow. It just sits back and waits, but once it figures out how these capabilities can be leveraged, they will definitely make use of them.” And so, don’t count states out yet.

Q: After years of intentionally not having a national digital privacy law, interest is heating up in the U.S., and what shape might this take? And has the U.S. lost the opportunity to dislodge Europe’s general data protection regulations?

On some levels I kind of question the premise of the question, which is, is there really growing interest in a privacy law in the United States? Is that like a ... Full disclosure, I’m Canadian. So nothing would make me happier, but that’s not really the approach the United States is taking. The United States has not really thought of a general privacy law. It’s had a sectoral approach. So we know there are privacy rules around health care information, there are privacy rules around education information that we here at Harvard are very much subject to. Have we moved from 1 percent to 3 percent, or 1 percent to 5 percent? But I don’t see the big outcry. Again that would make me thrilled, but I’m not sure I agree with it.

Let’s imagine that there is though. It is hard for me to imagine something ... I am not convinced that GDPR is actually all that effective, and I’m not convinced that it doesn’t have a whole bunch of perverse outcomes that I’m not sure the Europeans really intended or are happy about. I’ve heard some evidence already that suggests a lot of big companies actually like heavy regulated situations because it takes a lot of resources to comply with those regulations, and there’s been ... I feel like I’ve heard a couple of anecdotal stories that some alternatives to buying ads online have basically shut down because they couldn’t adhere to all the rules of GDPR, and so actually Google has done better in Europe. Which is probably not the intention. It depends what your goal was. Maybe if you want better privacy, that’s a good outcome. If you want more competition, it’s a bad outcome. And it’s not always clear to me which one the Europeans want.

So is GDPR effective, it’s not totally clear to me. I’m a little worried too that a lot of what’s happening with GDPR is like an earlier rule, which is why all these websites you go to have warnings about cookies, that I know everybody reads in thorough detail, and then really moderate their behavior based on that. No, they all just click the thing, and make it go away. I do think there probably is a certain degree of market failure here in that individuals’ ability to collectively act to protect, to create privacy outcomes that they’re interested in is pretty hard for me to imagine. So some sort of baseline law likely makes sense. But this is complicated space, and it’s hard for me to know what that looks like, and what will be effective, and then what will be the outcomes. Are we going to actually reinforce the status quo players, or are we going to create more competition which is a very, very different goal, and one that many people also share.

Q: I think at one point you mentioned the Facebook hearing, and the level of sophistication of the questions that were asked by members of Congress. Have you given any thought to doing some educating of members and staff of Congress on how they can more effectively deal with federal law, and oversight of the use of information technology by the federal government?

It does very much come to our minds. So one vehicle that I’ve had the pleasure of being involved in is ... Here at the Kennedy School, we have the new members of Congress conference that happens usually in December or January after an election, in which all new members of Congress are invited to come. And in the last cycle, I was asked to put together a panel on technology in policy. Now I want to be clear, this is limited time, and I just don’t want the question asker to feel like, “Okay, great this is good, problem solved.” We get a very, very finite amount of time. They have a very, very official schedule, but we are at least able to present a range of issues that they should be thinking about. My sense is that the solution here is probably less about educating members of Congress, and much more about doing two things, educating staffers of members of Congress because that actually ... it’s a shocking number of decisions get made at that level, and just the capacity is greater there, and it’s probably easier to reach them.

And then the other which I’ve taken a great interest in, and we have some fellows here at the Kennedy School who are working on this, which is to reopen ... What’s it? The Office of Technology Assessment, the OTA, which when Newt Gingrich was in Congress, they actually defunded and shut down. So this was kind of like the OMB, but was an office that really looked at technology, and tried to write reports, and create a baseline level of information for all members of Congress around technology issues. And we kind of killed it right at the very moment that all this stuff was starting to become important. So we have a Congress that doesn’t have an internal capacity like it does around budgeting. And so there’s a real movement afoot to recreate the OTA, refund that. And that strikes me as also an important piece of the puzzle.

Q: So what risks and opportunities do you see in applying technology such as artificial intelligence and blockchain?

Oh, yes, the blockchain. So let me deal with these very, very separately. Actually I wrote a piece recently that we can try to dig up that was called, “The Fast Follower Approach to Government.” It’s kind of my belief that at least on the government side, that we should use boring technologies. So let me just first kind of answer this selfishly. We’ve had the blockchain for 10 years now. It’s kind of incredible. And there is no universally accepted usage for the blockchain, where if you grab 300 CTOs from across America, or around the world, and you said, “Hey. What is blockchain for?” They’d all be like, “Oh, yeah definitely. This is the thing, best practice, completely acknowledged. We know this is the thing to do.” 

There is no accepted use case right now for the blockchain other than pumping and dumping coins in what are, I think, pretty shady markets. That’s not to say that there won’t be something, I’m just trying to put a really firm or harsh light on where we’re at right now. And with any new technology, what I tell my students is you think of a new technology like you’re entering a completely dark room. And you think there’s something valuable in that room, that the technology is going to help you with, but you don’t know what it is. What the venture capitalists are doing, is basically when they fund someone, they’re giving them a flashlight. And the degree of funding is how much light, and for how long. And then, maybe this is more your angel investors, but then you flash that flashlight around the room, and what you’re hoping to find is something of value that will allow you then to focus. And so then I say, well how much money have VCs poured in the blockchain over the last 10 years. Like one billion, two billion, three billion, probably somewhere in that range.

So we’ve poured three billion dollars’ worth of light into this dark room, and we haven’t found anything that’s obviously valuable. So at least on the public sector side ... I tell everybody, “If you’re in the public sector, and you’re funding something on the blockchain, defund right away.” Because why do you think your $500,000, or $1 million, or $10 million project is going to magically shine light somewhere in that room that the other three billion dollars’ worth of light isn’t looking at, and going to discover something and then also know how to use and maintain, and build something of value out of that. There’s no shortage of money getting put into that room, so why don’t you just let those people do it. And here’s the thing, is if they find something valuable, they’ll come and tell you. The problem actually is if they don’t find something valuable, they’ll probably come and tell you that they found something valuable anyways. So discerning what is a really valuable, that actually the skill that we need, not pumping more money into this.

So on the blockchain, it’s unclear that it’s useful for anything. I just haven’t seen the use case that does something better than what we already have where there aren’t all these unknowns that we should be concerned about, or haven’t figured out. I’m sure there are many people on the call who will either be surprised or disappointed, but I’m just not seeing it. And Bruce [Schneier], whose office is next door to me, said yesterday he saw some terrible app for West Virginia or something like that, where they’re doing some voting app on the blockchain. And he’s like, this violates the most basic rule which is ... He was like, “It creates no paper record, and the number one thing you want from a voting system is you always want to have a paper record. So you can run the election without any technology.” I’m a little bit perplexed by how popular it is. I’m not really, in the short term, worried about blockchain or its implications. 

I did get asked by a very senior in Silicon Valley on a trip out there with the dean, where he was like, “Aren’t you worried that the blockchain is going to displace the U.S. dollar?” And I was like, “Not at all.” And he was like, “Well, why not?” And I just said to him, “They’ll just make it illegal.” And he was like, “They can’t do that.” And I was like, “You’d be surprised what monopoly control over the judiciary and the use of force gives you.” And the dean, who was sitting next to me, so great, he leans forward and he said, “You know, in my lifetime, it was illegal to own gold bullion.” And that’s because FDR made it illegal during the Great Depression, because there was some concern that they had about ... and it was still illegal to own gold bullion until I think Ford. I think it was Ford who actually basically made it legal to own it. And I remember the VC’s mind was just blown. 

So if we think something is a threat to the monetary system, don’t doubt the power of the government to go and regulate, or make sure that it doesn’t threaten the monetary system. And I don’t understand why ... I don’t know what the benefit of the blockchain is versus existing currency. A permanent ledger does not strike me as something that actually most citizens care about, or there are all sorts of reasons why you would not want a permanent ledger of all the things that you spent money on. So I’m really, really not convinced that this is something that makes sense. 

The artificial intelligence strikes me as a completely different beat. Here there are demonstrated use cases of it being very, very high value. This is clearly a technology that is reshaping the way that we approach a number of sectors, and many of the questions we got are very concerned about algorithmic bias, using training data sets that are historical data sets with then of course bias against marginalized communities. So there are many, many, many more questions around this, a much more interesting sector, and one that deserves our attention.

Q: If you could maybe shine some light on what area in the public sector has digital technology had the least impact.

Well this comes back a little bit to the first point I talked about of these monolithic system versus platforms, and how really rethinking what is the technological structure of government, and I think there are huge opportunities in that, especially at the local level, but there are enormous capacity constraints. Most local governments don’t have the resources to even begin to think in these ways. Some of the big cities do, but your average city in America, this is just so far beyond what they would have the resources for. So we have a lot of work to do there. But you know, it really for me ... I think it’s easy to get obsessed with the technology. I gave a little bit of a rant on the blockchain there, you know, it’s very easy to grab a technology, and then say, “Let’s go apply it a bunch of places.” I’m actually not averse to that, I’m a little averse to public funds being spent to that. Let’s use private funds to do that. 

But a lot of the work that I think needs to be done is more of the do you know you’re in water, or do you not know you’re in water variety, which is less of the “Hey, how is A.I. going to save us? Or, how is this technology going to save us?” It’s actually about just the application of what is now, I would argue, relatively boring technology into spaces that continue to be paper based or worse. And I have one case study here at the Kennedy School that I teach a lot, which follows former Code for America fellows who, after leaving their fellowship, decide that they just want to keep doing this work, and serve in the public interest. And they go on food stamps to go experience the process of going on food stamps. And basically map out all of the ways in which that system is very, very challenging, for not just someone who is ... but they say nothing of someone who’s a new immigrant or homeless. And then go about rebuilding that system, understanding the medium in which we live, and really using not very ... without using the blockchain or A.I., just using really boring and normal web technologies but in a way that really focuses on who these users are, and what they need.

And that’s transformational work. That work needs to be done across almost everything that we’re doing in the public sector. And the capacity to do that is relatively limited, so this is not about what the newfangled technology is, it’s about really, how do we lift people up into understanding that they are fish in water, and what makes them skilled, and what makes them not skilled, and bringing them along.

Q: David, I’d like to ask you about micro-targeting of predictive analysis. Basically looking at what Cambridge Analytic mastered, the use of individual information to influence individual behavior. And how does government play a role to protect us from that.

Yeah, these are the big questions. So for those who ... I’m sure everybody understands the context of the question, but here is it. The issue is, are companies gathering enough data about us that they have a deep enough understanding about who we are, and what we care about. That you can then organize us into very, very small groups, and then target ads at us that will motivate us to behave in certain ways. And I do hear rumors of ad groups as small as 20 people, where you can find target groups so like they’re that big, and you will customize an ad relatively cheaply to that in an effort to ... In an election, often you’re trying to encourage those people to vote, or you’re trying to suppress the vote, and discourage them from voting. I’ll be honest, I don’t have great answers for you about how we should go about tackling this problem. In part because there are real free speech issues that come up in having government get involved. Part of those are tied by the way the Supreme Court has interpreted the funding of elections, which is deeply, deeply problematic. 

But there are also just more general free speech issues like what speech are you limiting people ... What are you limiting from people to be able to make on a platform like Facebook? And there may be a regulatory solution. My hope is that there is more likely going to be some form of self-regulation. And I say this because I think any regulatory solution probably won’t be adaptive enough. And I think it’s not unprecedented for us to think about a self-regulatory commission, resolution. And the story that I really glommed onto around this was coming out of the Shorenstein Center which here is the center that look at news media. There was a real concern when FDR was running for president that concentration of media power in the hands of some individuals was very, very high. So some individuals owned a lot of newspapers, and this is not quite micro-targeting but it is the concentration of power like you would see on a Facebook. And they were very worried about ... One of the owners of these said, “Listen, should I be using my network of newspapers to champion my perspective.” And particularly he was very anti-FDR. So should I be using that power, or should I not? And it led to I believe it was the Hutchins Commission, which was, what is the responsibility that newspapers have to the public? This kind of comes back to my public interest conversation. The publishers had a public interest conversation. So before the lawyers in the ’70s, you go back to the 1920s, the journalists had a conversation about this. And a set of, not rules but norms emerged, which, you know, there were things you could do, and things you couldn’t do, and you kind of kept the journalist side separate from the financial side of the newspaper, and so forth. And so part of me kind of wonders what the Hutchins Commission look like in the 21st century. I’m not the only person to ask that question. And what are the norms that we think we need the new media companies to live by, and then if they don’t live by them, then maybe we have to start using the stick of regulation. 

But I’m a little nervous about going there early because the regulation stick is pretty heavy, and I probably won’t get it right. And the lift is heavy, and if you get it wrong, changing is hard. So that for me is maybe the first step for us to take.

Q: Do you know of any potential solutions to policy challenges such as gun control or not filing your taxes that digital technology can help out with?

I think one of the dangers about being in my space is the belief that technology can depoliticize an issue. That there will be this technology solution that will change the calculus, and now everybody will be happy, or we’ll just solve the problem. And you just can’t get way from politics, that’s the harsh reality of it. As an example of this, there’s a great set of tweets that a professor had the other day, who went to Sears one day. She taught a course on the history of consumption, and she talked about how when Sears showed up, people don’t recognize now how revolutionary it was for African Americans particularly in the south. Before Sears, if you wanted to buy something, and you were ... The store was owned by the landlord, or the land owner, or the sharecropper. And they knew how much money you had, they knew how much credit you had, and they always served you last. You had to wait, it was a very dehumanizing experience, and then they might even deny you the right to buy, and be like, “I don’t think you have enough money.” They would just say no to you.

And the Sears catalog created freedom for those people, so much so, that there were these catalog burnings that took place, where white landowners would get together and grab all the Sears catalogs, and try to burn them all so nobody would have access as an effort to try to maintain control of this population. So there’s a wonderful story here about the emancipation or freedom that happens as a result of technology. In fact, that Sears had to educate ... Two things, this is great. Sears had to educate people about how to buy stamps directly from the post office because the store generally sold the stamps, and the stores refused to sell stamps so people couldn’t by from Sears. And so they did that directly. And then the white landlords starting to create these rumors that Sears was owned by African Americans to try to dissuade white people from using it. So then he did all these photos of him, and they’re like ... So here my point is, the story sounds like an emancipatory story, but there was a very real political battle.

There are universes in which Sears loses that battle. In which the landlords in the south get mobilized, and they pass laws at the state level that prohibit the purchasing of goods via the mail. That thing could have happened. So the technology did help, but there was a political battle that went with it. So I’m very, very cautious about ... kind of hoping that the technology is going to erase political battles. It probably won’t. It may reframe them, it may change the winners, and losers in them, but it’s probably not going to eliminate politics.

Q: A lot of the case studies we see around government using technology differently have been in service delivery area. Even the examples you were talking around about building better platforms for digital identity et cetera, the area certainly from my experience that I’ve seen that there has been less focus on how to use technology to improve the policymaking process either from the legislative side of it in terms of how elected officials get input and make decisions on legislative policy, to senior decision-makers in government, we still largely get briefings on a piece of eight and a half by 11 paper. Just curious if you could speak to that a little bit and give your insight on how, or how not, technology can be used to improve the policy part of government operations.

Great question, Ryan. I have written a fair bit on this, and I’ve been disappointed by the lack of change around this space. In some ways, technology has radically change the policymaking process. I’m willing to go on the record since I’m now speaking on the record, to say that in ways that nobody has really quantified, it would be really interesting to know how much do we think email has changed the policymaking process. Probably a lot more that we realize in ways that would be really interesting to quantify. Maybe a few thoughts on this. So the first is, the big challenge is not the technology, it’s that changing the mode of production is almost impossible. I do a whole class about why are we where we’re at, which is all focused around how the fundamental unit of information in government is the piece of paper, and once you build a whole infrastructure and decision making process around that, it is insanely hard to move off of that mode of production.

And that really is what digitization has meant in the last 20 years, has been the movement from a physical piece of paper to a Word document. Not changing the nature of the information, not changing the size and scope, but that’s about it. And I spend a lot of time thinking about how we could do things differently. And in Canada, they had this great experiment, or have this great experiment in GCpedia, which is an effort to create a wiki behind a firewall. And I always felt it was a huge opportunity to radically rethink how you made policy, having a lot more input from within the public service across the silos at an earlier stage before it ends up going to work before cabinet, and to PCO. That opportunity has just never materialized, and the reason it has not materialized is because the middle levels of management and government don’t want it to happen. They fundamentally see their job as controlling the flow of information.