Volume II Issue 11, November 2010
“New Media and Global Engagement”
By Ed Bice
Laila El-Haddad is a Palestinian from Hamas controlled Gaza City. She is also a journalist with a degree in public policy from the Kennedy School. To those who follow Laila’s writing, though, she is most prominently the mother of Yousef and Noor. I do not know Laila, but I‘follow’ her: @gazamom on Twitter. Last year, while traveling to Gaza, having secured the proper paperwork in person from the Egyptian Embassy in Washington, DC, Laila and her children, the youngest of whom was then 15 months old, were detained for 36 hours at the Cairo Airport before being told they would not be allowed to cross into Gaza because of the Israeli-Egyptian blockade. One agent stated, “We have no problems letting you live here [in the Cairo airport], really. We can set up a shelter for you. And no one will ever ask about you or know you exist.”
How wrong he was.
Laila, was not, as he might have guessed, using her mobile phone to text friends, she was tweeting her ordeal, including the words above from the customs agent, to the world in real-time. And @gazamom was trending.
This is new media.
New Media seems like a very good thing for the world, and at Meedan we are working to prove that there is a social upside to social technologies. However, as we consider what new media and the social web mean in the context of global engagement and citizen diplomacy, there are some rather daunting issues that surface.
All global challenges, from climate change to conflict, are in some way dependent on the question of whether the internet will ultimately increase the network diversity of information exchange or whether, given free choice to create our own channels and refine our information networks, we will evolve distribution structures that narrow our networks, and subsequently, narrow our thinking. There is no question that the internet and with it information, media, and knowledge exchange, are becoming ‘social’. What remains to be determined is whether this will lead to a world in which information exchange is more or less fundamentally diverse.
The most obvious and fundamental barrier to network diversity is language. While Arabic is the fifth most spoken language worldwide, ranking 27th in Wikipedia,, only about 1% of all content online is in Arabic. Success on improving digital literacy and communication is highly contingent upon access to content in Arabic sources.
The Meedan project was founded in a conference room at Carnegie Mellon University’s Language Technologies Institute (LTI). “What,” we asked, “would the world look like if we could wire the burgeoning social web for translation, using a motivated, distributed community of individuals to ‘crowdsource’ (use a distributed, ad hoc community) the translation of media, educational content, and conversations on the internet.” My friend Violetta Cavalli-Sforza (an LTI team member) suggested that ‘Meedan’- the Arabic word for a town square, field, or gathering place- would be a nice name for such a project.
At Meedan we source content from both Arabic and English and translate it into English and Arabic, respectively. We have a public facing news site at news.meedan.net with a team of 40 journalists and translators working seven days a week to curate and translate viewpoints from both the Arab and Western press; to date we have translated 6,615 articles and nearly 2,000,000 words. Our editors organize related excerpts and opinions from across the web providing a mosaic of diverse opinions for our web audience, which is composed of both Arabic and English readers, with slightly more readers coming from the MENA region. Our news.meedan team also publishes a round-up of translated excerpts from the Arab web on significant breaking stories for the Economist, and does live-blogging in two languages for partners such as the US Institute of Peace and the UN Alliance of Civilizations.
We are staffing professional translators to facilitate dialogue across Arabic and English between religious scholars for our inter-faith project with Cambridge University. With Qatar Foundation International we are implementing these same technologies in an educational setting, to enable collaboration between classrooms, teachers, and students. We are also working to translate Arabic language micro-finance loan requests on Kiva.org and English language Health and Science Wikipedia articles for a project supported by the King Abdul-Azziz City for Science and Technology in Riyadh, Saudi Arabia. Each of these projects presents unique application and interface design challenges, but common to all is the challenge of providing equal access to content, and, more significantly, equal voice, to a linguistically diverse user group. Scaling and cost are a significant issue for any translation based project. The technical approach we have taken is intended to promote incremental participation from a distributed network of contributors. All of Meedan’s content (user profiles, comments, excerpts, and full urls) is pre-processed through a Machine Translation (MT) engine developed over the course of our three-year partnership with IBM Research. Our paid translators correct this output directly on the application, and we feed all of the corrected output back into the engine to improve its accuracy. The twist is that any translated content that has not been ‘locked’ down by an editor can be revised, much like content on Wikipedia, by any authorized user on the system.
We are always pressed to talk about Machine Translation (MT), and so I will briefly address this obvious solution to the cost and scaling issues related to human translation. In the seven years that we have been working on this project, we have seen fairly dramatic improvement in the state of the art MT engines for translating documents between Arabic and English. However, it is critical that continued research energies be applied to this important language pair, particularly research that supports English to Arabic translation, which is currently under-served. While we have seen great progress with Statistical models for MT, these engines are only as good as the data sets that are available. And so, progress is dependent not only on continued research, but also on more enlightened policies and more innovative strategies with regard to generating open licensed data. As a first step, researchers and funding agencies should embrace and mandate open licensing for linguistic data generated from publicly funded research. In terms of strategies for scaling these datasets, the National Science Foundation is to be commended for supporting research that explores the internet as a platform for ‘crowdsourcing’ translation data. Precisely because MT will never completely replace the need for talented human translators (especially for scientific, literary, poetic, academic, and informal/ conversational content) efforts, such as Meedan that seek to leverage the talents of motivated individuals will remain a critical component of the multi-lingual internet.
In closing, I would like take up this theme of the individual and turn back to @gazamom. It is important that those of us who work in this space, and especially those policy- makers who are keen to promote new media as a vehicle for democracy building or other policy goals, be keenly aware that the power of new media rests in the social and personal nature of new media production. The tweets from @gazamom caused an upwelling of support because they were real, personal, and bound in all of the very personal facts that unfolded over 36 hours in detention. And, this is true of all the individuals who are interacting via social media platforms; our actions in these spaces are personal. When we practice social media, whether it be in the form of accepting a Linked- in connection, ‘liking’ a Facebook page, publishing a blog post, ‘favoriting’ a restaurant, or ‘following’ @gazamom on Twitter, we are establishing our personal digital identity, we are forming and sharing our online reputation.
This is profound in the context of citizen to citizen interactions–the hope is that there might be a new class of digital diplomats who are becoming more globally aware and forming networks that might amplify an important insight from a young scientist or answer a question from a high school teacher stuck on a lesson plan. This is the long, optimistic view of new media that could change the world.
However, through the same ‘social’ lens, we must recognize that there are relationships inherent between project funders and end users or practitioners. As policy makers consider how to best promote new media, they must not turn a blind eye to the perception that support for new media projects from governments or government affiliated funders fundamentally compromises the credibility of these projects. It is critical to recognize that for new media practitioners whose distribution channels are social and whose units of currency are reputation and authenticity, programs intended to support and promote their work can actually serve to undermine this effort. Further, such support may have an unintended chilling effect on adoption and use of new media tools more broadly in educational, scientific, and scholarly contexts. As a productive response to this consideration, I encourage funders and policy makers to support policies and research in support of the multi-lingual, social, and accessible internet, in support of an internet that allows us to simply understand what our global counterparts are saying.
ABOUT THE AUTHOR
Ed has published in national press including the NY Times, New Republic, and Mother Jones, and has been interviewed on national and international radio news programs like NPR, BBC, and CBS. Meedan was featured in the NY Times Magazine 2008 ‘Year in Ideas’ issue.As the chief architect of Meedan.net, Ed has been responsible for innovating technologies for cross-language social networking and crowdsourced translation.
Ed has been an invited speaker at the 2010 Harvard Advance Leadership Initiative, the 2009 UN Internet Governance Forum, and many other ICT4D events. Ed is a member of the Partners for a New Beginning (PNB) group, a council member of the Middle East Strategy Group (MESG) and is a co-chair of the United Palestinian Partnership (UPP). Ed has been an invited SBIR grant reviewer at the National Science Foundation in 2008 and 2010. Joi Ito included Ed in his 2008 book Freesouls, portraits of 296 people working to build the open web.
Ed and John Shore have co-authored a patent-pending approach to hybrid distributed natural language translation (HDNLT). Ed at- tended Carleton College where he received a B.A. in philosophy.When he is not online Ed spends time with his wife and two young children in Woodacre, California.