The Power of Crowdsourcing Knowledge in Online Communities

Faculty Insight

David Weinberger

Senior Researcher at the Berkman Klein Center for Internet and Society, Harvard University

Interview by Christina Nagler

In the age of the internet, online communities increasingly collaborate to answer the “unknown.” These campaigns of crowdsourced learning have ranged from the tech community’s 2007 satellite photo search for a man lost at sea, to the citizen scientists participating in Galaxy Zoo planetary discoveries.

David Weinberger

David Weinberger

According to David Weinberger, senior researcher at the Berkman Klein Center for Internet and Society at Harvard University and instructor of The Idea of the Internet, the power of crowds can be quite revelatory. In this interview, Weinberger shares insights on how online social networks have facilitated new discoveries and discusses the challenges of crowdsourcing knowledge.

Thanks to the internet, our role in mysteries has changed from spectator to participating sleuth. Do you think this is a viable trend?

There are some things that crowds can do very well. For example, they can find evidence across a large area. They can flip through evidence in a way that individual investigators simply cannot.

About 10 years ago, a well-known computer scientist, Jim Gray, did not return from a sailing trip. So his colleagues in Silicon Valley—computer scientists from Google, Amazon, Microsoft, and NASA—banded together to help find him.

They worked with DigitalGlobe to capture satellite photos of a very large area of the ocean, and made them available online. The entire web was allowed to go through the photos to see if they spotted any ships in the images. Unfortunately, they didn’t, but they well might have. That’s a very simple case of crowdsourcing in which people acted as sensors, using eyeballs and human judgment to turn up evidence.

But crowds can do something that I think is more interesting: they can engage a group intelligence that goes beyond simply turning up evidence.

There is a crowdsourcing science site, Galaxy Zoo, where the initial task was to go through hundreds of thousands of photographs of galaxies and classify them as being ovals or spirals.

In the case of the lost scientist, there was no facility provided where you talk with other people engaged in this activity. However, on this science website, participants could talk with others who were doing this pretty straightforward work of classifying galaxies.

But crowds can do something that I think is more interesting: they can engage a group intelligence that goes beyond simply turning up evidence.”

It turned out that some of these galaxies were neither oval nor spiral—some of them looked like green peas. And it was the crowd that noticed that. On the discussion boards, someone said, “Hey, I’m seeing something that looks like green peas” and someone else said, “Me too. I thought it was just me.”

Because the crowd network was allowed to talk, a new phenomenon was discovered that the scientist behind it had not been aware of. It was a significant step that went beyond simply using humans as sophisticated sensors, allowing them to develop knowledge on their own.

At sites like Reddit.com—at least at Reddit’s best—networks of strangers can develop knowledge in more interesting and complex ways through threaded conversations. They fact-check political assertions, question scientists, and float and then assess ideas about everything from gossip to medieval history to quantum theory. We now take this sort of networked knowledge for granted.

How do you think these networks are changing the way we interact with information and knowledge-building?

They advance and escalate something that we have always valued: engaging in knowledge development socially and through a network.

We are now seeing a move in our culture away from the idea that knowledge consists only of finished products that are filtered by experts, published on paper, and remain unchanged.

If you want to learn about something now, some information you can just look up and find out. But more interesting and important knowledge can be found by going out on the web and discovering these networks of people who are putting forth their ideas and arguing with one another.

Sometimes their differences are absolute, but oftentimes they are teasing out differences on a topic in which they generally agree. That is an enormously valuable activity that I think is changing our idea of what knowledge is and how it works.

In this more fluid environment of online dialogue, how are our notions of knowledge, truth, and facts changing?

I think the internet is uncovering a problem with our idea of what facts are supposed to do. That idea is relatively new in our history. It’s basically a nineteenth-century idea that a fact settles arguments—you cannot argue with facts.

In once sense, that is true. There are facts that the world is one way and not another. If human activity is causing climate change, then that is a fact.

But then we also use facts to mean statements of truth about the world. So, the thought was that we could use factual statements to settle arguments and to force agreements among people who disagree.

The internet is a global environment where people talk across cultural and topical domains, often without acknowledging that they are doing so. Lots of conversations go wrong that way, but they don’t have to.”

Senator Daniel Patrick Moynihan was famously quoted as saying that everybody is entitled to his own opinion but not his own facts, which is to say that if you have the facts you can shut other people up.

But it turns out that you can’t. There are facts about the world, but people are always going to feel free to disagree with statements about the facts.

And at this point, especially with the rise of fake news, I think it would be a healthy move for us to—and I want to be careful here—stop talking about facts, and instead talk about evidence.

The important thing about evidence is that, first of all, we can disagree about what the evidence is and what it means. That’s a very healthy human thing. That’s what scientists do.

Second of all, and this is crucial, we have to recognize that evidence is always in a system that has rules of evidence. Those rules are different in different systems. For instance, what counts as evidence in physics is very different from what counts as evidence in evolutionary biology or in a court of law or when you are arguing about the plot of Game of Thrones. In each of these cases, there are implicit rules of what constitutes evidence and how it works.

If you don’t keep in mind what those rules of evidence are within the domain, the conversation goes terribly wrong. It’s like a scientist and a person of faith arguing about the creation of the universe. Those are different domains with different rules of evidence.

But if you keep in mind that these domains have different rules of evidence and if you play by those rules, then you can actually advance the conversation.

I’m not saying facts don’t matter or it doesn’t matter who is right. It matters desperately who is right. I’m saying that the way to discern who is right and identifying the truth is through conversation that relies on evidence and rules of evidence.

The internet is a global environment where people talk across cultural and topical domains, often without acknowledging that they are doing so. Lots of conversations go wrong that way, but they don’t have to.

Share this story

More posts from Inside Extension


Add a comment

By submitting this form, you accept the Mollom privacy policy.