Channeling crowdsourcing into distributed work

How crowdsourcing affects government, research and the workforce.

I had an interesting chat recently with CrowdFlower founder and CrowdConf organizer Lukas Biewald about the growing influence of crowdsourcing on the worlds of government and research. Highlights from our conversation are included below.

Government agencies have had success with crowdsourcing in a number of areas, from Be A Martian to using grand challenges in open government. What other crowdsourcing applications could government employ?

Lukas Biewald: Consider the work of Ushahidi, or the work that my company did in Haiti translating text messages. How else can you provide such rapid access to potential volunteers? I also love the way that CrisisCommons involves people in crisis relief who want to help but don’t know how.

You can also look at The Guardian’s example, where they set up the means for people to go through the British government’s expense reports.

You can see the Sunlight Foundation using volunteers in Transparency Corps to go through campaign data, like filtering fliers in tagging projects.

Think about DARPA’s Grand Challenges. They got so much more than $1 million of research out of that investment. Major institutions worked on it, along with many amateurs. And then there’s also the XPrize & Energy Department, which are teaming up on a 100 MPG car. [Winners of the energy efficient auto contest were announced yesterday.]

In the video below, Biewald joins Brian Herbert of Ushahidi, Robert Munro of FrontlineSMS and Leila Janah of Samasource to talk about how they deployed a critical emergency communications system after the earthquake in Haiti:

What does crowdsourcing offer academic researchers?

LB: This is a great way of doing research. Almost every grad student in the Stanford linguistics department is using crowdsourcing platforms for their studies. This is a really useful tool for social scientists or researchers in other areas to have access to data or tools they didn’t have before.

There’s also a whole genre of new research that is coming about through crowdsourcing efforts. For example, why does WIkipedia work where other models didn’t? This work leaves such a big digital trail. There’s a lot of data out there for people who are interested.

What are some of the issues around crowdsourcing?

LB: There are big questions here. The trend toward crowdsourcing work can cause a big drop in wages in some industries. It’s outsourcing on steroids. Before, it could be hard for a startup to outsource a big part of its business. Now, firms like 99 Designs have moved almost all of their design work offshore. Is that fair?

What’s happening there is happening across industries. Companies are outsourcing legal work, e-discovery, advertising and more. But this raises questions: When people do things for free, does that really work? When you pay in virtual currencies or reputations, how does that change things?

This interview was condensed and edited.

Editor’s Note (9/17/10, 9 p.m): Portions of this post were rewritten and refocused.


tags: ,
  • Crowdsourcing is interesting, but what is more important is getting our scientists to work together in shared codebases and repositories.

    I start the AI chapter of my book with the following question: Imagine 1,000 people, broken up into groups of five, working on two hundred separate encyclopedias, versus that same number of people working on one encyclopedia? Which one will be the best? This sounds like a silly analogy when described in the context of an encyclopedia, but it is exactly what is going on in artificial intelligence (AI) research today.

    Today, the research community has not adopted free software and shared codebases sufficiently. For example, I believe there are more than enough PhDs today working on computer vision, but there are 200+ different codebases plus countless proprietary ones. Simply put, there is no computer vision codebase with critical mass.

    A big part of the problem is that C and C++ have not been retired. These languages make it hard for programmers to work together, even if they wanted to. There are all sorts of inefficiencies of time, from learning the archane rules about these ungainly languages, to the fact that libraries often use their own string classes, synchronization primitives, error handling schemes, etc.

    It is easier to write a specialized and custom computer vision library in C/C++ than to integrate OpenCV, the most popular free computer vision engine. OpenCV defines an entire world, down to the Matrix class so it cannot just plug into whatever code you already have. Meanwhile, if you want to write your own specialized computer vision library, you don’t have to start from scratch as there are many great libraries for graphics, i/o and math. There is plenty of quality free software for building your own computer vision library, but the OpenCV library is in C/C++, so we haven’t moved beyond this first stage.

    To facilitate cooperation, I recommend Python. Python is usable by PhDs and 8 year olds and it is a productive, free, reliable and rich language. Linux and Python are a big part of what we need. That gives a huge and growing baseline, but we have to choose to use it.

  • Thanks for the fascinating interview.

    I don’t agree that crowdsourcing is properly described as “outscourcing on steroids,” as Biewald suggests. You could just as easily say it’s consulting on steroids, when you think of the “expert on-demand” side of crowdsourcing (like what we do at In other cases, like the work of Ushahidi (and crisis-mapping), it’s work that can’t be done in any other way. What looks most like outsourcing, and might raise the kind of concerns Biewald describes, is mechanical turk-type platforms that essentially trade in “human intelligence” piecework.

    @hypios we’ve also written on ushahidi and the brief history of crisis mapping:

    and on the uses of open data in government.