Abstract

America is losing its social cohesion.  We this as the combined result of (i) the loss of a shared reality or basis for discussion; (ii) confirmation bias that supports fringe views; (iii) frictionless publication that grants equal time to propaganda and misinformation, and (iv) a political under-commitment of the American population.  We propose a cross-platform (web, android, iOS) system app called "Enlightened" that addresses this by adding a second dimension beyond popularity to rate information distributed through social networks.  We call this measure “credibility.”  Ideally, people will use this to pass along articles and statements that are both significant and come from sources who are “credible.”  Its measure is a function of how knowledgeable and reliable the information transmitter is, I.e, an author gains credibility points by demonstrably exposing themselves to alternate views, engagement in active discussion, and validation of their post by fact-checking.  To the extent that people use this measure, they may virally become encouraged to re-distribute more honest information and allow society to gradually reach some convergence and compromise on issues.


Introduction

Social scientists have long analyzed and decried deep divisions in American (and global) societies. Boorstin considers this in 1962 (The Image) and attributes it to the demands of continuous media and the concomitant news cycle. Putnam (Bowling Alone, 2000) attributes much of this to lack of active participation in public affairs and events such as bowling night.  More recently, Cramer and Hochschild have studied the divisions between urban and rural America, and between Southern Tea Party members and the more liberal  North.  In the latter two author’s work one message screams loudly: a large segment of the population feels powerless, disrespected, and marginalized. We summarize this as humiliation, and we suspect that humiliation is a strong political motivator. While this is not a new phenomenon, it is certainly one that we all sense is more exaggerated now.


Some attribute this to a general polarization that has been progressing for many years as a result of many media-irrelevant technologies. For example, veteran politicians attribute it to the airplane — members of congress return home rather than fraternize. This reduces their familiarity with other views and other people. In the 1990’s Massachusetts Congressman Joe Moakley reinforced this view in a private conversation with the author: he mourned the lack of comity among politicians simply because after a day of vigorous argument they no longer went out to play golf together. The Internet and in particular social media are also blamed by dint of the frictionless way that fringe views can get wide circulation and thereby become self-reinforcing (refs…).  Whereas once we have limited news media that was both popular and centrist, we now have cacophony and no critical mass around which we can rally.  When there are few outlets, they all become similar, as noted by Hoteling in 1929.


Another consideration is more mainstream media such as talk radio and broadcast TV networks such as Fox and MSNBC.  Their profusion has obsolesced Hoteling and allowed minority and bizarre views to remain vibrant. In 35 years of research into the invention of open media at the Media Lab at MIT, we can observe that media in general is non-rivalrous.  More media does not detract from existing media; even when it competes with it, they both can easily survive. Fox, for example, was founded in 1986 but its growth accelerated in the mid-90’s coincident with the consumerization of the Internet and the transformation of chat rooms to Internet fora. (Ref)


The short summarization of this is that there is a general sentiment that we are retreating into local bubbles that include only those who share our a priori views.  While this is a general truism it appears that the walls of that bubble have gotten notably firmer and they risk transforming us from a society into an archipelago of local (whether physical or virtual) views and actions.  We call this a lack of social cohesion and we think this potentially is more significant than “false news”.  In other words, it is not the veracity of what we hear that is the problem.  Rather is the receptiveness we show to it that is dangerous.


The importance of this phenomenon has long been recognized.  Lippmann considered it in 1922 in “Public Opinion,” with reference to news, propaganda and a public open to manipulation.  Hoffer address “The Nature of Mass Movements” in the 1950’s in the context of World War II.  He offers counterintuitive reasons and observations on who is susceptible to such movements.  In particular, even in the era when Christ was alive, his religious movement was more effective in cities where people had lost social cohesion and felt alienated.  It was far less impactful in the countryside among paisans, now pagans, because they had social structures that served them well enough to inoculate against a new radicalism.


The question we address is whether Internet media can be built to combat the divisiveness that distorts public discussion and reduces social cohesion.  In particular, we attempt to turn the frictionless nature of modern publication into a feature that can elevate discussion rather than debase it.  Further, we attempt to do this in a way that we call “viral.”  That simply means it can propagate incrementally rather than require advance commitment by a majority.  It is successful at any scale.


System Design

We address the loss of social cohesion by using a “divide and conquer” approach. First, we look for individual solutions to the problem sources to define requirements for our system. Then, we design an application that meets these requirements and incorporates both viral and gamification elements. Lastly, we consider deployment issues and motivate the need for a new platform.


Addressing the Loss of Shared Reality

The lack of social cohesion is partly caused by the loss of shared reality. People in their bubbles generally surround themselves with people with similar opinions and consume biased media that resonates with their prior views. However, they often do not inform themselves through other sources and usually do not expose themselves to different views or debates. Furthermore, media bias exacerbates this problem. Hence, each bubble has its own reality and over time, the shared reality between these bubbles might be marginalized.


What can we do to restore shared reality? Figure [lenses & bridges] illustrates two approaches to this problem. The first approach is to broaden people’s views by exposing them to a diversity of views. The second approach is to create channels that allow information flow between the bubbles – allowing people with different takes on reality to share their view. While both approaches are steps in the right direction, they both raise a couple of concerns in terms of practicality.

The first approach raises following questions: Why would a liberal person be interested to hear a conservative view and vice-versa? Why would I take time to read a story from multiple perspectives? Why would I switch my current platforms for news? We can reformulate these questions as requirements.




  • System Objectives Recap

    • Solutions for each problem:

      • (i) Present a broader view / [“Breaking the bubble”]

        • How to do that?

          • Intro -

            • Broadcast News are typically structured into {as discussed in a Group Meeting}

              • What/Where/When/Who happened? [Green - Fact]

              • Why [Yellow - Factual context]

              • Anything else [Red - hypotheses, opinions, base-less conclusions]

          • Present AllSides.com

          • Present Video Summaries

            • Generated by Superglue

        • What are other options?

        • Why not other options?

        • Concerns

          • Just presenting different views will harden the position

      • (ii)

        • Disfluency disrupts the confirmation bias (Hernandez et al., 2013)

        • Need to incorporate a mechanism to do that

          • E.g.: Tests (“Test My Knowledge”)

          • E.g.: Game - “Defend the opposite position”<— a way to increase your “influence points”

        • What are other options?

        • Why not other options?

      • (iii)

        • Re-introduction of friction using “credibility” measure.

        • “Knowledge Points” characterizes the “credibility” of someone who disseminates information

        • What are other options?

        • Why not other options?

        • Why is “credibility”-only not enough?

          • Given a “credible” person that shares your political views, this person can be abused for excuses etc.

          • This measure can easily be criticized and not taken seriously

      • (iv)

        • Introduction of “influence points” to incentivize commitment

        • What are other options?

        • Why not other options?

    • Challenge: how to bundle individual solutions together and pack it into an app / platform extension

  • Platform Choice

    • Why are existing social platforms not suitable?

      • Case Facebook

      • Case Twitter

      • Pew Research Center - findings regarding news consumption on social media

    • Why are existing technical solutions not suitable?

      • Google Chrome Extension

      • Redundant: Facebook App / Twitter App

    • Thus, we propose a new system: “Enlightened” (web/mobile app)

  • Design Considerations

    • End-User Perspective

      • What are user habits -> Pew Research Center

      • What are drivers for users to use “Enlightened”

    • User Experience

  • System Description

System Implementation

  • System Overview

    • SuperGlue

      • Create Video Summarization Module

        • Video Summary based on CC

        • For starters, 5 most important topics of the day

          • 3 video summaries from left/middle/right sources based on topic

      • Create Test Generation Module

      • Create Discussion Topic Module

    • Mobile

      • Front-End

      • Back-End

    • Web

      • Front-End

      • Back-End

Deployment

  • Strategy

  • Channels

Post-Deployment Evaluation

  • Post-Deployment Analysis

  • needs it go viral

  • Why did it work?

  • Why didn’t it work?


Outlook

  • What needs to be done?

Bibliography