How the Kennedy Campaign Used Tech and Data

[music]

Announcer: Listener-supported WNYC Studios.

[music]

Brian Lehrer: It’s The Brian Lehrer Show on WNYC. Good morning again, everyone. We would like to think that we're more than the sum of our data. In 2020, we take for granted that advertisements and our news feeds pretty much know our interests better than we do. For some people, that might be part of the appeal of social media or online shopping apps. The more data you supply, every Facebook like or Amazon purchase, the more these companies are able to predict what you're going to pay attention to. To others, that's concerning. There are ethical questions. Questions over privacy.

Of course, we've seen companies and countries use data to manipulate voters and influence elections. None of this began with Facebook, Amazon, Google, or Twitter. My next guest traces it all the way back to the 1950s and '60s to one company called the Simulmatics Corporation, and to their People Machine, a computer program designed to predict and manipulate human behavior. Jill Lepore is professor of American history at Harvard, and a staff writer at The New Yorker. She's the author now of If Then: How the Simulmatics Corporation Invented the Future, and she's with me now. Welcome back to the show, Professor Lepore, always an honor.

Jill Lepore: Hey, thanks for having me, Brian.

Brian: First of all, am I saying the name of the company right? Is it Simulmatics?

Jill: Simulmatics.

Brian: Simulmatics. You write that the story of Simulmatics would be more comforting if the people behind it were villains. If they weren't villains, who were they?

Jill: [chuckles] Maybe it would be even more fun to write. It is a fascinating story to come across. I think in the end, more interesting that they were incredibly well-intentioned social scientists, most of them, led by a very idealistic liberal from New York, a guy named Edward Greenfield, who in his own advertising agency, the Edward L. Greenfield company. [unintelligible 00:02:26] really Madison Avenue Advertising Agency. Extremely charismatic guy. Also, intellectually quite ambitious. Really curious about the latest research going on in universities. A real connector guy who seemed to seemingly knew everyone.

He wasn't a very staunch Democrat and was involved in Adlai Stevenson's presidential campaign. In 1956, he worked on local campaigns. He worked for [unintelligible 00:02:58] in '52. He was really frustrated by the Democrats foot-dragging on civil rights and was keen to think about a way to convince the Democratic Party and in particular, Stevenson, to take a stronger position on civil rights.

Somehow this led him to the work of social scientists who were studying voting behavior in a really rigorous quantitative way for the first time using mainframe computers, which were brand new, and had the idea with a number of these scientists to try to devise a program that could simulate an election, and that could be the thing that they could sell to the Democratic Party in hopes of changing the direction of the party.

Brian: Interesting. Tell us about the influence of Simulmatics and the People Machine on the election of JFK. He was one of their first big clients.

Jill: Yes. Greenfield founded a spin-off of his advertising agency in 1959 called the Simulmatics Corporation in order to seek a contract with the Democratic National Committee to pay for the initial research, which was putting together the People Machine, compiling all the data that would be needed, which really public opinion polls and census data and coming up with a mathematical model for voting behavior using this cutting edge research that was being done a lot of it at Columbia, in fact.

They did an initial study for the DNC that they claimed was quite influential in 1960, in shaping the platform at the Democratic National Committee, which took an incredibly strong position on civil rights of the Democratic Party, which really had been the party of Southern-like conservatives for a long time. Then when Kennedy secured the nomination, they got a contract with the Kennedy campaign to produce three more reports which they delivered to the campaign late August, early September of 1960. Everything that they told the Kennedy campaign to do, the Kennedy campaign did, and Kennedy won. Greenfield, who was a great publicity guy, put out the word that Simulmatics was responsible for his victory.

Brian: Fascinating. Was Kennedy apprehensive about the technology at all around the ethical implications of gathering data and using it to manipulate people? Or it was more like such early days for this kind of thing that he was all gee-whiz?

Jill: I don't know what Kennedy himself got. I wasn't able to come across anything directly from Kennedy about it. It became a public relations problem for the incoming president when Simulmatics took credit for his victory because Kennedy had campaigned. He made a big issue of automation, which was hot but in the same way, we’re always debating the future of work right now. Automation in the 1950s had been an extremely [unintelligible 00:05:56] political issue. People were extraordinarily anxious about automation in the Democratic Party. The party of Labor had taken a strong position for job security, job retraining. Kennedy had given them the speeches about that before he came out that his campaign had been run by a machine. It was a problem and so the campaign denied ever having used the services of the Simulmatics Corporation.

Now, Pierre Salinger, the famous pressman of the Kennedy campaign, issued a rebuttal when the story broke. There was a bit of a flap up about it, and you get the sense that-- Kennedy was willing to use the machine, but they didn't think that the company would use them. [chuckles] [unintelligible 00:06:42] that if you're going to use this technology to aid your campaign, you should probably commission it some kind of promise of non-disclosure afterwards.

Other people were really worried. The reason it became such a flap up is when the story broke that the campaign had employed Simulmatics. Newspapers across the country, and there's editorials and every newspaper saying, "Oh, this is a very bad idea. This is going to destroy American politics." People were deeply concerned about the lack of responsiveness to voters as voters about mechanization.

It seemed like taking democracy for granted. There were concerns, of course, about the leadership of a president who would take his cues from a machine. They're not that different from-- They're an extension of the concerns about polling. Should you really ask a poll before you decide what to do as a president? Actually, that's not what you're elected to do. It's not a meaningful enough measure of the role of the people because there's many problems just going inwards as well.

Brian: Of course, that's a complicated conversation because you want to know what public opinion is and you want good data on what people really think in a democracy and then apply your own leadership standards to that as well. I think that's a complicated conversation. On the prediction of people's behavior in the context of politics, and in the context of 1960s society, which is what we're talking about so far, we'll bring it up to the present, but that was a time of racial unrest. There were protests in cities around the country, and you write that Simulmatics believe their technology could predict race riots.

Then you say cite James Farmer, activist with the Congress of Racial Equality, who said that instead of trying to predict riots, people should be doing something to address the problems that led to the riots. What were the consequences of trying to predict race riots? Why did they use a computer to try that at all?

Jill: That actually is quite complicated because, again, the first report that Simulmatics prepared for the Democratic National Committee was called the Negro voter in the door. It's in fact and people forget that public opinion companies tended to not ask questions of Black voters in the 1940s, 1950s. Then Gallup started doing his public opinion polling in 1935.

Gallup pretty much decided never to poll Black voters because in the South they couldn't vote. In the North, he thought there weren't enough of them. Also because his column was syndicated in newspapers across the country and southern papers threatened to drop the column if he ever asked people questions about civil rights or if he reported the political views of African-Americans.

There was a huge gap in knowledge [unintelligible 00:09:41] among the many protesting accuracies of polling for much of the 20th century is that it really is the public opinion of white people. The guys who've come to Simulmatics really objected to that. Their report for the Democratic Party-- Some of the very first people to really take seriously the Black vote. It had a Black Votes Matter commitment.

They did this initial report to say, "Here's what the Black vote in the North would be if you took a stronger position on civil rights and here are the swing states the Democrats could win," that Kennedy could pick up in what was one of the closest elections in American history. Here's how that we could take the White House back for the Democratic Party after eight years of Eisenhower would be taking a stronger position on civil rights because we would be able to pull these Black voters in the North away from the Republican Party.

There's a real commitment to racial justice there. That said, you have to wonder, why did it take that for the Democratic Party to budge? We're talking 1960s [unintelligible 00:10:44] Greensboro. Once counter citizens are gone, they're citizens across the South. It's not a mystery what the Black voter wants. It becomes a weird construction of the white Northern liberal intellectual. "I wonder what the Black voter wants. We could build a machine to tell us. We could use our extraordinary artificial intelligence to devise a [crosstalk]."

Brian: Rather than have relationships.

Jill: Right. We live in a world that is built that way. Where we now take for granted that political candidates and officeholders will segment the electorate and you would figure out the predictive behavior of voters by minute demographic bodies and send us particular messages around it but this was new then. People thought even for cultures to do this was extraordinarily ethically compromised. What happens over the course of the 1960s is that this machine that was built-- I would think that the founders of Simulmatics and Greenfield would have essentially [unintelligible 00:11:49] help enfranchise the Black electorate, was then really moved into the world of predicting political resistance and counterinsurgency.

Before the Simulmatics Corporation was hired by the Kerner Commission in 1968, that's the Johnson National Commission on racial civil disorders after the long summit after the riots in Newark and Detroit. The Kerner Commission hired Simulmatics to study race riots and they do a lot of work to think about whether there are tools to predict race riots. [crosstalk]

Brian: You see a connection--

Jill: [unintelligible 00:12:27] .

Brian: Go ahead. Finish your thought.

Jill: I think the connection there is not from the early voting work. The connection in the prediction of race riots is from the work that companies like Simulmatics do in Vietnam or [unintelligible 00:12:43]. The way the Department of Defense thinks about countering communist insurgency by figuring out what the hotspots would be. Figuring out what representatives of the population are likely to be persuadable by communist messages and therefore, how could they be persuaded to reject those messages?

All this work that's done in a psychological warfare in Vietnam, it's pretty consistent with how politics run by machine works as well. You come up with a mathematical model of the population, you measure its opinion, you figure out the right message to send to a particular targeted population in order to get the opinion results that you're looking for. One of the things you're doing is looking for signs of unrest so that you can see which population needs to be addressed. The work in the US, not just by Simulmatics, but it's a forerunner of predictive policing.

The city of Detroit assembled a big computer database in the 1960s as well in order to try to predict racial unrest using crime data and population data, housing data. This was new then and far more controversial then. It also worked infinitely less well than it works now. [[chuckles] It works in an unfair way now, but there was not enough data to meaningfully make these predictions at the time and the computers people were using were so much slower, obviously.

Brian: You mentioned Vietnam and you write that the US war in Vietnam was the first war waged by computer. That went well. Getting back to race, you see a connection between trying to predict race riots back then in the 1960s and today's reliance on cameras and data by police departments. Can we draw a line between Simulmatics' 1950s, '60s, and more aggressive policing in cities today?

Jill: I think you can. Certainly by no means the main piece of what Simulmatics was up to, but it was very much the sensibility of that fascination with predictive programs in the 1960s that leads us to where we are now. It's in many ways, the lack of a public rejection of an objection to these programs that I think we just to where we are now.

The reason this story is important to me and why I wanted to tell it is, we often think that we are in this wholly new era and that Facebook or Amazon or Spotify, this algorithmic world that wouldn't have it, came out of nowhere. It just poured out of the heads of these Silicon Valley geniuses and then mixed together with piles and piles of money and now the rivers run through across the country and flows in the blood of our vein, but it didn't come out of nowhere. Really, it came out of the Cold War. Most of this stuff is based on the tools of psychological warfare that involve collecting information about your enemy, sending messages that can manipulate their behavior, and extract a different outcome from their actions than the one that they were initially taking.

That's what this behavioral science when it marries computer science is trying to do both in the Department of Defense and with the commercial application toward political campaigns and just your basic. Simulmatics had contracts with Colgate-Palmolive and Ralston Purina, [unintelligible 00:16:33] to try to get people to switch brands. That is all a piece. I think it's helpful in reckoning with what the machine looks like today, this People Machine that we're now all trapped in. It's useful to figure out who built it and when to give something that seems to have no past.

Brian: Hey, can I take a short detour to your latest New Yorker article? Because I thought that was very intriguing. The headline Is Staying In Staying Safe? That, of course, is a big assumption that a lot of people are making in this year of coronavirus. The subhead says Indoor life has its dangers, too, but building-design specialists have big plans for us. What's the heart of that?

Jill: There's a whole healthy building movement that's part of the world of architecture and design that very much predates this pandemic. Things like beautifully, decorated, well-lighted wide stairways in buildings that encourage people to use the stairs instead of the elevators or an escalator. Many buildings, especially residential buildings have these tiny narrow back of the building stairwells that are dark and tend to be not well-repaired, but for better healthy living, we should encourage people to use stairs if possible, that kind of thing.

There's a lot of concerns about ventilation. It's the case that we think much more about pollution as an outdoor phenomenon, but the writers of the books [unintelligible 00:18:13] I was reviewing for this essay, make the point that we spend most of our lives indoors, very much more than 90% of our lives indoors in the United States, long before the pandemic and now we spend much more of our lives indoors.

Most of us experience pollution that's outside. We [unintelligible 00:18:29] indoors, but it never seeps into our indoor spaces and maybe doesn't have good ways of getting out. There's a lot in this literature and in this research about lighting and fresh air and now all these things seem infinitely more urgent as things like school buildings try to deal with the maintenance issues of very poor ventilation systems that have not been kept up and have not been updated. I'd say it's reckoning with some of that research.

Brian: As we're seeing intensely debated this week in New York City school buildings around town. Let's take a phone call getting back to "If Then". Kevin in Bridgeport, you're on WNYC with Jill Lepore. Hi, Kevin.

Kevin: Hi Brian and Jill. I was just fascinated by your discussion and it seemed to me that the idea that these are Cold War or they started in the Cold War, it's like we're in a Cold War with social media the way, presumably, Russian Intelligence or bots that are trying to stir up civil unrest and pull apart. It seems like they're utilizing the knowledge or the tactics or some element of what you're discussing there with social control. I don't know if I'm using the right word, but just the way that you're describing this, it seems like there's a parallel.

Jill: Yes, that's a really interesting point, Kevin. Thanks for that. In some ways, and it's the inverse of that moment when the scientists that I'm talking about who go on to work for Simulmatics, they engage in projects that are designed-- As you know the US involved during the Cold War and really trying to influence election, say, in the course of politics in, say, Latin America.

There's a project that is a computer simulation of Latin American political systems and economies that's proposed. It's called Project Camelot. A lot of these guys working on that project in the early 1960s. It's a secret project that was exposed and Latin American legislation say, "Who do you think you are simulating our economy and simulating our political system in order to influence our political behavior and our economic development?"

The Johnson administration has to shut down the program but those programs continue, just that are hidden and through other sources of funding. The Simulmatics Corporation, for [unintelligible 00:21:11], takes the project to stimulate the economy of Venezuela. The reason you have [unintelligible 00:21:17] an interesting intellectual project, [unintelligible 00:21:20] could you simulate an economy and then see where you could make an intervention, but by what right? It is the kind of meddling. One of the things that emerges from these projects in the 1960s is a really powerful ethical debate about it and especially, of course--

Brian: I apologize that my phone is ringing in the background. I told you never to call me here. Just kidding. Go ahead.

Jill: [laughs] [unintelligible 00:21:49] anguish of Vietnam calls all this into question. There's a huge debate. The anti-war movement is really focused on the research that's being done on university campuses to advance the war effort, not just the work of people in the biological sciences, developing and the chemical sciences developing new agents of warfare, but people in the social sciences who are involved in advancing the project of the Vietnam War on university campuses and university buildings with University faculties. There's a big movement to divest from that work and a real ethical reckoning with it. Yet, all those mechanisms that you use now in Silicon Valley on anyone who has an iPhone.

Brian: One of the central figures in your book is Ithiel de Sola Pool, who was in your words, a numbers guy who taught at MIT and walked the halls of the Pentagon.

Jill: He was a quite prominent defense intellectual and I think very dedicated and hardworking, brilliant social scientist. He believed in the end of the Cold War and in the containment of communism and he believes [sound cut]

Brian: Whoops, did we lose Jill's line just there? I think we have to rehook with Jill Lepore for the last few minutes of our conversation about her new book, If Then: How the Simulmatics Corporation Invented the Future. In the meantime, I'm going to go to another caller. Phillip in Astoria. You're on WNYC. Hi, Philip.

Phillip: Hi. Thanks for taking my call. I wanted to point out that I think there's an underlying lapse that really has to be faced and isn't really. That is that social scientists and cognitive psychologists and politicians and advertisers understand that humanity in general doesn't have a psychology, it has a mechanics which is completely predictable and manipulable.

Also, that extends not just on a mass scale but to individual scale as the micro algorithms that the major ten companies used has proven. At the same time that we have this knowledge, so to say, in the world of those who know, we have a general cultural myth in forming ideas that we're individuals with an identity, with ideas that were self-actualized. We have a [inaudible 00:24:37] and that's the illusion, you have a mechanic. The reason this is upsetting with [inaudible 00:24:46] what's happening with all these [inaudible 00:24:48] puts the why to be true to the notion that we are individuals and are special, and we're not. We're hard and we're predictable. There's a real rub there between what is and what we believe as a culture and [inaudible 00:25:04].

Brian: I'm going to leave it there, Philip, and get Jill Lepore, who's back with us now to respond. Jill, I know you couldn't hear the beginning of that but the caller is saying, a deep reality underlying this is that we live in the illusion as human beings that we're individuals when really we're parts of herds and the fact that we can be manipulated in the ways that Simulmatics and other data manipulators can do, demonstrates that. That may be beyond the scope of your book, but there it is.

Jill: [laughs] First, I think somehow I've offended the technological overlords because my phone-

[laughter]

-[unintelligible 00:25:50] in the middle of talking about the failure to reckon with the consequences of Vietnam. I understand that and I hear that frustration about this feeling of powerlessness that I think many of us have and is exacerbated in this moment and we're all the more tied to our devices as the only way to interact with the world. I, myself, don't believe that.

I also think it's important to recall that our political systems have a representative constitutional democracy, depends on a wholly different understanding of the human condition. That we have to be individuals who are capable of making decisions for the public interest in the interest of the common good and that those are rational choices. This is an enlightenment notion of how human beings act and are, rather than a protestant predestination, Calvinistic notion. That's the one that I subscribe to and I think it would be hard to get through the day otherwise.

I do think is, like I said, it's an interesting point because that sense of powerlessness is very dangerous. It leaves us with less ability and less strength and determination to act as individuals and to reach out to one another and to believe that we can somehow escape this experience of being wholly manipulated.

Brian: I'm glad that the--

Jill: I think we can. I don't think we belong to herds and tribes and it's inevitable that we act this way.

Brian: I'm glad that the technological overlords were ambivalent enough about cutting you off that they let you back on the phone because that was a pretty profound way to end moving from the nature of machines, which is a lot of what we were talking about, to something basic about the nature of humanity, prompted by that caller. We will leave it there for today with Jill Lepore, professor of American history at Harvard and staff writer at The New Yorker. Her new book is If Then: How the Simulmatics Corporation Invented the Future.

Jill will be participating in some live Zoom events to talk more about the book if you're interested. One is tonight as part of the Chicago Humanities Festival. That's 7:00 PM Central Time if you're there so eight o'clock Eastern Time. You can register on their website at chicagohumanities.org. Look for the "events" link at chicagohumanities.org. Tomorrow she'll be speaking with Danah Boyd of Data & Society at the Brooklyn Public Library at 7:00 PM Eastern Time. That's also a virtual event. You're not going to the Brooklyn library, right?

Jill: [laughs] If only we were.

Brian: You can register for that at B-K-L-Y-N. You know how they shorten Brooklyn. B-K-L-Y-Nlibrary.org. We'll post links to these and others on our show page at wnyc.org. Click on “Brian Lehrer Show.” Jill, thank you so much.

Jill: Thanks so much.

 

 

Copyright © 2020 New York Public Radio. All rights reserved. Visit our website terms of use at www.wnyc.org for further information.

New York Public Radio transcripts are created on a rush deadline, often by contractors. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of New York Public Radio’s programming is the audio record.