The science of going viral

A higher ed digital communications expert develops a computer that mimics Jon Stewart
By: | Issue: March, 2017
February 16, 2017

Andrew Lippman is one of the foremost experts on viral communication and digital life. As a founder of MIT’s Media Lab, Lippman had been studying this field long before many of us ever heard of the internet.

Currently, he heads the lab’s Viral Communications Group, which explores how infrastructure-free, personal communications systems will transform society and businesses. He explores how communication has become embedded in our daily lives, and how social networks can be reshaped to work as fluidly as face-to-face conversations.

“Viral ideas are grassroots innovations that start at the edges of networks and industries, take only small investments to begin, yet build through social and technical opportunism to swamp entrenched, vertically integrated companies,” he says.

Lippman will discuss some of the Viral Communications Group’s work as a keynote speaker at the UBTech 2017 conference in June.

Let’s start by talking about the Viral Communications Group.

We started the idea in 2002 to study viral communications systems and how they scale. We began with thinking about radio, but soon generalized and expanded it into cellphones and the like. The more cellphones and radios you have, the more congested things are and the less actual communication you get done.

The question we wanted to explore was, is there a solution to that? Are there ways to make things like networks scalable—where the networks get better as they get bigger—as opposed to getting overloaded as they get bigger?

That’s what you mean by viral?

Right. We mean systems that can start small—decentralized systems, peer-to-peer systems, democratized systems—where they can scale and grow without limits, but at the same time where there’s value added to them as they grow.

Give me an example.

The archetypical viral example is the spreadsheet. The spreadsheet was not a top-down, centralized, large-scale program brought to you by IBM. It was started in an apartment, and it started small.

It percolated its way through a small company that used it to test ideas such as mergers and acquisitions, then others started working with it, and it gradually became the standard.

The interesting thing was, as the spreadsheet grew, the users themselves made it better. Spreadsheets got macros, which meant that the users could actually add to the idea of what a spreadsheet is, and the macros could propagate along with it.

That’s the way we think of viral systems. In the Viral Communications Group we study publication systems and news systems and broadcasting systems. In a sense, we’re the peer-to-peer corner of the media laboratory.

How do you apply this to things like the news, for example?

Well, people have done a lot of thinking about this since the election, right? The legitimacy of news and the legitimacy of media has been called into question. The apparent ease with which you can generate false news is suddenly on people’s agenda.

Whatever anybody thought about news before the campaign, they are thinking about it differently since then.

By the way, this isn’t new. We were interested in this stuff 40 years ago. I still have a thesis that one of my students did in 1986. She considered what the impact of electronic news reading would be.

This was before the internet really, and before social media, when we were thinking about electronic news—reading the news on a screen instead of a hard-printed newspaper. She explored what happens when you lose the context of the newspaper, for the news.

How so?

When you talk about anything like news, there’s the raw information, but there’s also a lot of context in it. Whether it’s a headline, or whether it’s above the fold, what kind of type they used, whether there is a photograph, whether it’s in The New York Times—all those things contribute to context.

Much of that context threatened to be lost in the world of electronic media. She wrote that in 1986, but she could have written it last month and it would be just as important.

Pretty much what everybody thinks is a prime example of news these days is The Daily Show. Every night, [former host] Jon Stewart would take one particular story and he would dig into the background of that story.

He would pull up clips that clarified or pointed out foibles of what someone said today compared with what they said six months ago. You don’t get that in the world of modern journalism.

How does the Viral Communications Group approach that?

We first had to ask, how did he do it? The answer is, he had an army of interns and an army of DVRs. They can go through the various broadcasts about a particular topic, and they could frame that story for him to do that night.

Stewart had another advantage that the rest of the news world doesn’t have—he had the leisure to pick his story. If there was a story that he couldn’t expand in an interesting way, he could skip it. The traditional broadcast news shows can’t do that. They have a responsibility to report on everything that happened.

So, we built a really big DVR, and we built some computer programs that behaved a bit like Jon Stewart’s interns, and we called it SuperGlue. The idea was to come up with a searchable, usable archive of broadcast TV news.

You can see it from different perspectives. You could lace together different elements of stories to see how they evolved. We looked at stories, and we looked at how they evolved, and we have demonstrations showing that from different perspectives.

SuperGlue is running today?

Yes, it’s watching 19 television channels and recording them and dividing them up, giving you tabulations of what’s said and so on. Now, it’s becoming apparent that the narrative associated with the news may be more important than the facts that underlie it.

It’s not just a matter of finding more facts, it’s a matter of also understanding the context of how facts are used in the world and how those facts influence people.

Another way of looking at it: You might replace the word “news” with “propaganda,” and say maybe some of what’s wrong today is there are things that are neither false nor true, but are used to deliberately misinform—often for political ends. That has undermined people’s faith in the media.

There are lots of people in the world of news who do that. It’s important to understand the behavioral aspect of it.

What do you mean by behavioral aspects?

There’s a seminal paper in this area, written in 1952 by a couple of psychologists. It’s called “They Saw A Game.” The paper was written about the last football game of the season in 1951 between Princeton and Dartmouth. It was a really rough game, wit many penalties.

Two psychologists, one from Dartmouth and one from Princeton, interviewed people who had seen or knew about the game in the months after the game actually occurred.

The conclusion was they saw two different games. The Dartmouth people invariably decided Princeton instigated the penalty behaviors and the Princeton people decided Dartmouth started it.

What does that teach you? Well, one of the things it teaches is that you and someone else could look at the same underlying facts and come up with completely different perspectives.

You discard facts that don’t agree with your social position or worldview, and you accept facts that reinforce that kind of thing. That’s the behavioral aspect.

It seems to me that anyone who is working on news has to think more about the behavioral and psychological aspects of it, and the context socially—as opposed to the context totally within the medium itself—as to how you proceed thinking about news.

There was a time when the most trusted man in America was Walter Cronkite. Now we seem to have gone completely the other way, and the least trusted voice in America is the voice of legitimate media.

Tim Goral is senior editor of UB.