i-054913da66b2d8ca014bab963ea3594f-mit audience.jpg

CAMBRIDGE, MASS. — After the announcement of the Knight News Challenge winners came the first “plenary session” at the Future of News and Civic Media Conference. The topic is “Crowdbuilding,” with the following panelists:

Chris Csikszentmihalyi is director of the MIT Center for Future Civic Media.

Gabriella Coleman is an anthropologist who studies the ethics of online collaboration.

Karim Lakhani is an assistant professor at Harvard Business School who studies distributed innovation systems.

Chris Csikszentmihalyi, MIT: The term crowdsourcing has been to the detriment of the news business. “Debian” is a collaborative GNU/Linus release that is powerful enough for space shuttles.

Coleman: Debian founded in 1993, with thousands of pieces of software. Has more people working on it than any free software. There’s a “social contract,” a “constitution” and “free software guidelines.” [Plays clip about following the rules of the constitution.]

Openness is about transparency and code but it’s not open to all participation. People think hackers are free-wheeling libertarians, but with any large software projects, their problem-solving skills go into institution building. The membership enter a gateway that all developers must pass through, which is three steps and is a long process that takes months. You might be hooked up with a German mentor who makes you answer 60 questions. They check a new person’s identity and checks on technical expertise.

i-8b7047ddec5ecd90a5b39af636448916-crowdbuilding.jpg
Gabriella Coleman and Karim Lakhani

So why was this created? The project followed a crowd logic. Debian became really popular in the late 1990s, so there was an influx of new developers. So people wanted to include commercial software to start competing with other Linux developers. That disturbed the older members, so they stopped accepting members for more than one year.

[Shows email showing all the new requirements for developers.]

We can reach clarity on how to build stable and ethical institutions from the ground up, using this as an example. One of the pitfalls from Debian and Wikipedia is that it’s difficult to have open-endedness and freedom but still have rules and red tape. Virtual projects are also marked by apathy as well as flame wars. With projects that are demanding, flame wars come up constantly.

Journalism has always had a strong ethical element. Clay Shirky has argued that the Internet has facilitated group creation. And he’s right. It’s mind-boggling. He treats it as a net political gain, but I think it has a downside. With the proliferation of so many groups you’re competing with members and there’s a problem with fragmentation. Every group doesn’t need to do everything. There’s a limit to the proliferation of project. Collaboration is more important than ever before.

Q: How do you get people involved in projects?

Coleman: Some people are only slightly interested but then get ethically on board. With Debian you don’t have to be into the ethical questions to get involved. Projects should be interesting on their own outside of the ethics.

Lakhani, Harvard Business School: Two brothers from Indiana created a Super Bowl ad for Doritos, from a contest that had 1,900 entries. Two ads from the contest were in Top 10 USA Today rankings. Now this has become the regular ad strategy for Pepsi.

Joy’s Law haunts innovation efforts. “No matter who you are, most of the smartest people work somewhere else.” — Bill Joy, Sun Microsystems

Here’s an example at MIT: Bob Langer, who is very smart and is the expert for tissue engineering. The fact is that the nature of knowledge is that it’s highly distributed — it’s unevenly distributed in society. Eric von Hippel said, “Knowledge is sticky.”

You start with 1,900 ideas, and then they get filtered and funneled down to the best ideas. We want “extreme value,” the one killer idea, not the average value of everything. Most organizations are struggling with how to get distributed knowledge and then figure out how to filter them. You can do it through competition with a diversity of approaches. Contributions tend to be substitutes, as in the Apple app store.

With collaboration, you can get contributions that range from mix & match to co-productions. Can be driven by intrinsic motivations.

Out in the world, there are firms doing collaborations and contests. Here’s a company that’s doing it very well: Threadless. The idea is to design a T-shirt and become famous. You upload your design, and then they get voted on, from 0 to 5.

Threadless community: 500,000 members; 800 designs per week submitted; 55,000 people have submitted design; 50 million votes made.

[Shows video of Threadless leaders explain how it works.]

Lakhani: The core work for the company is being done outside the firm. And even choosing the designs comes from the community. Out of 800 new designs, they only make 7 each week. So it’s very hard to win. Knowledge is out there, so you can enable a set of customers to generate and vote on the best ideas.

Why do people participate? Most people in a contest will lose. It’s non-rational behavior. There’s a range of motivations: They feel creative; it’s fun, building a skill; doing what you want to do; increasing their knowledge. Open source people volunteer an average of 14 hours per week of work.

Q: For people who just got a News Challenge grant, how could you pare down lessons learned to share with them?

Coleman: It depends on the nature of the project. There are all sorts of virtual projects that have innovated in interesting ways. There’s a need out there to connect projects, either technically or not technically. I don’t see a lot of groups that exist to do that.

Lakhani: There should be a push for transparency. You have a self-archiving history of the project. Put everything on a public email list. With Apache Group, you almost need to show people “candies” or bite-sized things they can accomplish. You will get power distribution, where fewer people do most of the work. You need to just accept that. Competitions and open source often focus on results and not credentials.

Q: What about incentives? If you have $10,000 for a contest, how do you spread that around?

Lakhani: I’m all for allowing for smaller prizes. NASA had a contest where winners got to see a shuttle launch.

Demos of MIT Center for Future Civic Media Projects

Now comes some demos of Center researcher projects. First up is Rick Borovoy, who is showing off Homeless Neighbors. Talks about seeing people sell homeless paper called “Spare Change.”

Borovoy: So we wanted to build home pages for the sellers of the papers. So we put a sticker on the paper, which points to the seller’s home page. Sticker says, “Peaches is your Spare Change News Vendor. Check out her home page!” [Shows her home page.]

There’s also a ChipIn widget that has helped her raise $160 so far for a career development program. Someone tweeted her page and it got a couple hundred hits on it. The sticker is probably the most transformative part about this. Changed her relationship to customers, it personalizes it. This week, the vendors agreed on a question to ask people, and there’s an online poll.

We wanted to build a community around the vendors.

Next project: LostInBoston.org. One thing about Boston is that it’s hard to find everything. So I wanted to do something about this. What if we could crowdsource better signs and get the government to put up those signs. I went to government and got a pretty tepid response.

But I was walking around Kendall Square and saw lots of private property signs, so maybe there’s an opportunity there. Lo and behold, it started to work. Got a pilot project going with the Mass. College of Art. We got a sign for the Avenue of the Arts, and put it up there, on private property.

i-4919b2ee129a4e044f3cb7cd0b47c293-junkyard jumbotron.jpg

Now we’re going to crowdsource suggestions for future signs. And the government is now interested in helping us.

Next up is the Junkyard Jumbotron. How can you create a spectacle without a big cost? Our idea was to stitch a bunch of laptops together in an ad hoc display. Takes a picture of the array of laptops and then email it to the special address. This is a live demo with eight laptops.

The system then parses it, and sends the image to all the laptops.