Creating Technology for Social Change

MIT Ignite Talks: A Civic Media Smorgasbord

Watch live streaming video from knightfoundation at

Mitch Resnick, director of the Media Lab’s Lifelong Kindergarten group, introduces the panel by describing the culture of the Center for Civic Media and the Media Lab at large. There isn’t a central planning committee that assigns projects to the students and researchers. Instead, the Center’s work is based more on prototypes and explorations—some of which fail, but some of which become wildly successful. “The key to success is diversity and iteration,” says Mitch. He gives a disclaimer: each of the projects being presented are at different stages of the iterative process.

The panel is organized into 5-minute “Ignite” talks—the slides will automatically advance, whether or not the speakers want to. To kick off the panel is Matt Stempeck. His project LazyTruth seeks to fight misinformation in our email inboxes. “Our brains aren’t necessarily designed for empirical truth… word of mouth is our most trusted source.” We trust the “real people” in our lives over editors and authority figures.

When misinformation lands in our inbox, we have to always be vigilant and seek the truth. Matt claims, “The truth is out there, but we need a bridge to get to it.” Often more difficult than finding the truth is the delicate art of telling people that they’re wrong. LazyTruth is like watching a YouTube within an email—it waits in the background silently, drawing in content from Factcheck, Politifact, and Snopes to debunk virus myths or political rumors. This type of technology is particularly important in 2012, an election year. We can study how these chain emails mutate and change over time. We can perform network analysis and fight against the spread of misinformation.

Next up is Chris Peterson, a graduate student in the Comparative Media Studies department. He starts by talking about social bookmarking site Digg and the Digg Patriots, a group of users who would strategically repurpose the site’s algorithms to promote certain political messages. This phenomenon, which Chris calls “self-generated censorship”, has appeared in many social media sites, including Digg, Facebook, and Reddit.

The same algorithms that make social network sites valuable, make them susceptible to subversion. Adding users to the feedback loop can be a double-edged sword. When we talk about internet censorship, we usually think about large institutions exercising central authority, such as governments (China) or corporations (Google). However, with self-generated censorship, we see the silencing of opinion from a decentralized space.

For example, when Facebook blocked j30strike, it did not only block the site itself, but any page that linked to the site, preventing all opportunity of conversation around the censorship. The systems are being gamed, and most people will never know that it is happening, as it occurs invisibly.

Now we have Nathan Matias, who brings up Michigan lawmaker Lisa Brown, who was silenced for using the word vagina during a debate. He shows us a graph that indicates that the number of women journalists has dramatically decreased over the past few years. The technology we have for keeping track of these statistics are vastly limited. Nathan has built some software to mine this data, determining who’s writing and who’s being written about, based on U.S. Census data on baby names.

His algorithm looks at the use of pronouns to determine whether articles are written about males by males, about females by females, or mixed gender. His preliminary findings, looking at the New York Times, show the following: 67% of articles are totally male, 26% totally female, while 7% experience the “Pat problem” (gender ambiguous). Female writers tend to be more gender-inclusive. In the summer, articles are more gender-inclusive but tend to lack bylines, while in the winter, they are more gender polarized.

Pablo Rey is next, presenting PageOneX, an online open source distributed analysis tool that looks at newspaper front pages and highlights them in a visual way. The tool also allows users to superimpose Tweets atop the front page visualizations, to get a fuller picture of the social conversation. According to Pablo, information flows rapidly between mass media and social networking sites; with some stories, like with the Occupy movement, stories get lost and never make it to the front page. Pablo wants to study these cases.

He has looked at big new stories and resultant information flows—the Japanese Tsunami, Arab Spring, and War in Libya—to understand how they are being represented in the media ecosystem. He wants to incorporate transcripts from television and radio, online newspapers, as well as Twitter and more social networking sites. The idea is to be able to customize which sources to collect from, to provide an in-depth look at each media outlet.

Matt and Nathan return to the stage to present a collaborative project—Media Meter. Matt brings up a Jezebel story that compared packaged foods to the images on the packaging. Unlike food, media consumption can be tricky. It’s harder to tell if we are consuming a “balanced” media diet. Media Meter compares different news sources with each other and compares them over time. The platform aspires to support individual goals and track personal information diets.

Nathan begins to explain the technical aspects of the Media Meter. He processed 20 years of New York Times articles to see the balance between U.S. and World articles. Media Meter utilizes Amazon’s Mechanical Turk to process articles, as well as hand-crafted modules and plugins. Data is fed into technologies like the Berkman Center’s Media Cloud and Pugg. Matt and Nathan invite the audience to provide data, ideas, and insights to enhance their project.

Rahul Bhargava takes the stage, wanting to talk about data visualization and broadening the definition. He brings up Scott McCloud’s Understanding Comics as a landmark text in visual communication. The information value of different visual forms depends hugely on audience and context. Rahul’s project Data Therapy is focused on providing community groups with visualizations for those unfamiliar with the visual language of data visualization.

Working with community partners, Rahul still wrestles with the question: how do we know what technologies/forms to use? Visual displays of information is hard, because it requires knowledge from so many fields. We have to make use of what they already have in their toolkits—for example, crayons, Legos, or Microsoft Excel. Rahul tells us to think about the thinking process that these community groups might be using.

Their goals aren’t just about presenting data, but also reinforcing the community and the values that they want to achieve. “It’s an empowerment story,” Rahul says. One of the concepts that Data Therapy is playing with is that of the data mural—a format that does not only present information, but invites further reflection and thought on that information.

Charlie DeTar begins his presentation by describing the twinkle (jazz hands, spirit fingers) and how it’s become a de facto gesture of approval, particularly in the Occupy movement. Chief among its successful features was structure—a structure that reinforces equality. In the 1970s, feminists started describing structureless meetings as being “tyrannical”. Charlie asks, how do we bring horizontal structure online and scale democratic participation?

The truth is, communication works drastically differently online than it does in real life. We need to design a consensus tool specifically for the web? He looks at three key components of online consensus:

  1. Understanding. Looking at Reddit’s scoring algorithm, it’s designed to find out what’s hot, but not build consensus.
  2. How do we structure online to match offline groups? Groups matter and should be at the center of the process.
  3. Participation. The great paradox is that if they don’t use it, it’s not democratic. How can we design this tool with a participatory design process? How do we involve community groups? Charlie says we have to educate users offline as well as online.

Huan Sun is next, describing NGO 2.0, focused on empowering Chinese grassroots NGOs in central and western China. What kind of challenges are they facing? The two major areas are communication and resources. For the first, Huan proposes to build NGOs’ social media literacy. Facebook and Twitter may be blogged, but there are many alternatives, such as Weibo. Every year NGO 2.0 holds Web 2.0 Training Workshops, to train NGOs’ digital literacy. So far 134 NGOs have learned how to integrate social media into their initiative, starting a snowball effect. All the class materials are also available online for smaller NGOs.

Addressing the second challenge, Huan identifies corporate social responsibility as the key. Corporations want to diversify their social giving; furthermore, state-run projects have been the subject of multiple scandals over the past year. NGO 2.0 matches NGOs to corporations by location and interest areas; it also invites the public to participate in the discussion. Huan says that building ICT projects is not as important as focusing on offline initiatives. NGO 2.0’s two main areas of focus are co-design and hackathons, specifically optimized for the Chinese non-profit space.

Paulo Rogério, a Fullbright Humphrey Fellow from Brazil, is the next speaker. Brazil has the 2nd largest black population in the world, but they face challenges such as poverty, a lack of political participation, and mis- or underrepresentation in the media. Hate speech and human rights violations are commonplace, in part due to the tight relationship between politicians and media organizations.

The solution is to empower black communities with the power of social media, allowing them to represent themselves. Paulo’s project Midia Etnica seeks to develop online, cross-platform tools that increase participation of black youths in the civic sphere. This is not only the case in Brazil, but all across Latin America, and Paulo hopes to expand the project in the future. He is optimistic about the many civic media projects going on around the Center.

Becky Hurwitz is up next. She describes New York’s recently-passed Domestic Workers Bill of Rights and the various educational initiatives that the Center is exploring to publicize the changes. For the first time, domestic workers are now covered by the same labor laws as other workers. Similar initiatives are being explored in California, and even internationally.

The system was built upon VoIP Drupal—a plugin for the open-source content management system Drupal built by fellow Civic Media researcher Leo Burd—and allows domestic workers to post audio clips via their phones. This is accessible by any kind of mobile phone—smartphones, “dumbphones”, and even landlines.

Now Leo himself takes the stage to talk about the What’s Up initiative. He starts his talk by describing South Woods, a community in Wisconsin, faced with unemployment and poverty—37% of the community was not connected to the Internet. The What’s Up system allows community members to stay informed of local events (employment, food, education, fun) through both online and offline means.

For example, signs are located at the local Walmart, Job Center, and Food Pantry. A paper bulletin is also put out, distributed at multiple locations. Finally, a community hotline has been set up, that allows community members to learn about events and contact event organizers directly. Leo reflects upon the lessons learned: chiefly, community partners—in this case the Incourage Community Foundation—are essential in any community initiative. Leo looks ahead to the future: at the local level, he wants to promote local adoption of the What’s Up system and improve its technologies; in future, he hopes to expand the program to other communities.

The next talk starts with a clip of Obama talking about the deportation of immigrant youths and students—dreamers. In a voiceover, CMS student Rogelio Lopez describes how these undocumented youth have gained new media literacies and taken to creating videos, websites, and other transmedia productions to empower other students and tell the stories of their lives.

Dan Schultz, recent graduate of the Media Lab, takes us on a tour of his many projects. ThoughtBox is a time capsule of ideas; Distributary allows users to create and share auto-curated G+ Circles and Twitter lists; TierRaid hopes to explain the world through ranking systems; Carstragram allows you to take a photo from your car by simply pointing and shouting; Ambiartist allows digital artists to create artwork to portray what goes on online; Media Empire is a real-time strategy game in which you gain resources by reading online; (Ink) Droplet allows users to store small bits of data in a physical “droplet”, like a bookmark; SocOpera allows users to take information from social networks and convert it into soap opera form; TLD(R) explores the corporatization of top-level domains; QR TV is a TV timestamp platform that uses QR codes to enhance the television experience; the Meta Meta Project is a tool that extracts metadata from various media; ATTN-SPAN (“Because watching paint dry can be more fun”) embeds C-SPAN videos in relevant news articles; News Your Own Adventure (which later became NewsQuest, see below) mashes up news articles about the same topic to make it possible for readers to decide what angle of story to dive in; WallPaper uses proximity to determine what information to show where; the GPF is a community hub that allows content to appear in more than one place; Newsjack lets you hijack the frontpage of any website; Truth Goggles, Dan’s master’s thesis, is an “automated bullshit detector for the Internet”.

Now Dan Sinker takes the stage to tell us about the OpenNews hackathon that preceded the conference. Inviting twenty of the conference attendees and opening the Media Lab’s doors to forty programmers, data experts, and journalists, Dan organized a hackathon around data & storytelling. Out of the eleven final projects, Dan presents five:

  1. Surfbored, a channel-surfing experience for the web.
  2. NewsQuest turns news consumption into a choose your own adventure. It allows users to enter a URL, aggregates different news sources, but only shows the first paragraph. Based on questions it narrows down the list and then shows the next paragraph, etc.
  3. Condition of anonymity, exploring the reasons the New York Times has given for granting anonymity to a source.
  4. The Outside Mappers Guild takes Instagram photos and allows users to clump them together topically—possibly an invaluable resource for news organizations.
  5. NewsDiffs monitors top news outlets and keeps track of changes made to articles.

Dan says hackathons aren’t about building code, they’re about building communities. You hope that interesting code will come out, but what you really want is for a long-lasting community to emerge. In this case? We got both.