Clay Shirky: online crowds aren't always wise

Clay Shirky, leading commentator on internet technologies and author of Here Comes Everybody, last night backed away from his earlier enthusiasm for the online wisdom of crowds in democratic decision-making. He suggested that Government use of social media should focus more on “small groups of smart people arguing with each other”, than national-scale engagement online.

We’ll, that’s my interpretation. You can listen for yourself … and find more on Twitter from last night and today’s at ICA.

A few years back Clay said that the ability of groups to organise online and challenge conventional engagement was “the glory of this medium”. He now believes we need more checks and balances.

When speaking at LSE, he was asked how he would now advise government on the use of social media. Earlier he had highlighted how a group pressing for legalisation of the medical use of marijuana had made this a prioritised item to be addessed by the new President Barak Obama, on his official transition website (see video above).

I would not be concentrating right now on the kind of large legitimating moves …. precisely because of the hijack model because in a way, even with new tools … tightly interested groups have a way of throwing issues higher up the charts  … I would be worrying instead about how to get good ideas out of small groups.

If you want to know where new interesting useful idea are going to come from, don’t look at crowds and don’t look at individuals, look at small groups of smart people arguing with each other. Historically that’s been a big source of change, whether you are talking about the Invisible College or the French Impressionists. Instead of having Government scale, or social scale initiatives, kind of have your say stuff  …

(Here Clay breaks off to quote an early negative example of online crowds how UpMyStreet – mainly devoted to postcode-level activity – was taken over by racist ranting when offering national-level discussion. I found this post by one of those who developed the UpMyStreet site via David Brake’s blog. David there cites Clay’s earlier views that occasions when online polls come up with results that are unrepresentative  are part of “the glory of this medium”).

Clay went on to say of national-level discussions that these problems are hard, bordering on intractable. He suggests (I assume to Government) putting together small groups of people who have some common appetite, for example networking ombudsmen in government departments as a community of practice to share what they hear from interactions with the public.

We are not ready for “massive legitmating moves of unstructured participation across the larger issues. That’s the first time I’ve said that in public.
I have for years believed that, and now I find myself saying that if it is all about a group of potheads trying to gain a position I’m less interested than I was.

Clay’s views may get some attention in UK Govenment because the question was asked from the front row by Cabinet Office Civil Servant William Perrin.

Since the question was about what advice Clay would give to Government, his remarks are as interesting in the context of Government online public engagement, as they are in the legitamacy of online campaigning.

Charlie Beckett, director of the journalism and society think tank Polis, chaired the discussion, and here underlines the issue of legitimacy raised by Clay.

But how do you distinguish between the campaign by Mysociety against MPs who tried to cover up their expense claims, with a bunch of potheads trying to get their spliff decriminalised? In Clay’s words, we “need to find an algorithm that works”.

Here’s the link on the MySociety campaign, where Tom Steinberg blogged:

The vote on concealing MPs’ expenses has been cancelled by the government!
In other words – we won!
This is a huge victory not just for transparency, it’s a bellwether for a change in the way politics works. There’s no such thing as a good day to bury bad news any more, the Internet has seen to that.

Judith Townend has an interview at online journalism news:

Shirky says he previously made certain assumptions about the result of what he calls ‘crowd wisdom’ and its positive impact for democracy. Now he believes that public pressure via the internet could be ‘just another implementation layer for special interest groups’.

It is ‘not fair’, he adds. There is a need to redress the political checks and balances in place in order to control the influence of such groups, he explains.

For example, during the Obama campaign, he watched the campaign for legalisation of the medicinal use of marijuana become a prioritised item on the website.

But, Shirky explains, while this type of online phenomenon is a ‘net positive’ for democracy, it is not ‘an absolute positive.’ It doesn’t necessarily mean these representational tools are a replacement for the vote, he adds.

“Are we really going to let a Digg-style voting algorithm commit the federal government top issue to the ‘wrestling with medicinal use of marijuana’?” he asks.

Shirky hopes to open up the debate on this issue: “There needs to be some mechanism by which executive or legislative branches can say we are taking this under advisement, but we are not taking dictation [from special interest groups].”

“It’s clear that it’s yet another environment in which special interest groups have to have some kind of check and balance against them,” Shirky says.

He doesn’t challenge the value of the mySociety campaign on MPs expenses: it alerted people to what was happening.

“It was classic news cycle timing,” he says, referring to what some said was a government tactic of a ‘good day to bury bad news’.
“What MySociety did is break that cycle and and publish it in media that doesn’t have a cycle. There is a new mechanism, in addition to referendum and political representation, which is not the same as casting a vote,” he adds.
“It is democracy in action, at such a young stage that we don’t even know how to integrate it into the rest of the democratic mix.”

Benjamin Ellis has a long and thoughtful analysis of the presentation: Mass collaboration – snow joke:

Clearly mass collaboration isn’t going to solve every problem. For the first time in public, Clay said, “I don’t think the technology is ready for the mass legitimisation of initiatives… …There need to be checks and balances applied”. That is a big, and wise, shift from his previously utopian view of what could be achieved. I’ve posted about crowds not providing the wisest answer for every situation before. When we think about the idea of direct access into the political process, we might want to think carefully about what exactly we are wishing for. The tools are fantastic for gathering feedback and generating content, but decision making requires a degree of sophistication that the tools do not provide, yet.

Update: full video of Clay at LSE here

Previously, Clay Shirky on:


  • February 4, 2009 - 3:38 pm | Permalink

    Thanks for this detailed summary David. In a way, this is not really large or surprising shift from a previous position for Shirky. All of the examples and arguments in ‘Here Comes Everybody’ are about bottom-up, user-generated activity / sharing / co-ordination and such large scale ‘legitimising’ or democratising initiatives which he distances himself from here are largely top-down invites for ‘users’ to generate content and activity. This feels more like a refinement and a clarification than any great shift of emphasis.

  • February 4, 2009 - 6:11 pm | Permalink

    There is discussion of this on the e-democracy researchers forum where I am looking to data to support the check and balance that the promotion of real identity would give you:

    Also, my experience with local-up Issues Forums where we’ve designed for accountability and effectiveness, that the wisdom of crowds can and is emerging. See my neighbourhood:

    I think is fair to suggest that what we need is a patchwork of local spaces and that starting nationally may be the least effective way to turn online talk into community building action.

    Steven Clift

  • Pingback: Bookmarks for February 2nd through February 7th | DavePress

  • February 8, 2009 - 9:41 am | Permalink

    Many thanks, David, for this excellent summary on a fascinating topic. I have always thought that the wisdom of the crowd phrase is a bit misleading. For me, it is tempting to take it to mean that whatever the largest number of people vote for must be the right answer, but of course when it is put that bluntly it no longer seems so appealing! I would see it as meaning: the more people you bring into a debate the more likely you are to get the fullest possible picture of solutions. This is particularly useful when there is a fairly clear right answer (e.g what is causing the bug in this piece of software?). It is also good at promoting innovation. If there is a consensus view that there are 5 possible solutions, then involving a lot more people is likely to generate unconventional solutions from people who would not normally get the chance to put their ideas on the table. But then if you vote on all the ideas (even with a representative group of voters), there is no particular reason to think this will generate the “wisest” decision.

  • Pingback: Socialreporter | Clay Shirky at LSE: full video and audio

  • February 9, 2009 - 7:38 am | Permalink

    The fact that crowds aren’t always wise shouldn’t really come at a surprise to anyone, especially when you consider that the groups that formed during the various e-participation efforts on don’t even fit Surowiecki’s definition.

    There was never any commitment on part of the transition team to let the Digg-style voting algorithm determine what Obama’s priorities should be. Instead, the Citizen’s Briefing Book was intended solely as a compilation of “facts and recommendations to be considered while crafting and enacting policies.” Nothing binding in there, just another channel for public input.

    Finally, it’s debatable whether what we saw on really amounts to decision making. I’d argue it was some low-level form of general input gathering (with ranked preferences for questions and ideas), but on the actual issues no choices were made.

  • Pingback: Clay Shirky on and crowd wisdom at Intellitics

  • James
    February 10, 2009 - 11:32 pm | Permalink

    Though crowdsourcing may not always be the way to find answers to biggest human kind problems, it’s been used since long time ago: It´s called democracy.

    Now online, crowdsourcing *is* always working with a small group of people, so Shirky may need to check his numbers on this. Last time I checked, top proposals submitted at had between 10,000-12,000 votes. Statistically, that is a small group of people, considering that the US alone has approximately near 250,000,000 internet users.

    So, the problem may not be in the answers, but in the questions. What are you asking the crowds? What are you going to do with that?

  • Pingback: Here Comes Everybody - Mobilising Tube Commuters | Travel experience!

  • Pingback: FrontlineSMS:Medic | Our team’s first face to face meeting

  • Comments are closed.