SXSW: The Wisdom of Crowds
Speaker: James Surowiecki - New Yorker columnist
Eugenics founder story that’s in book. Francis Galton at fair in England. People were asked to guess the weight of the ox after it was slaughtered and dressed. Big crowd - some experts (butchers), most not experts. Galton took all guesses and applied statistical analysis. Turned out that crowds guess was right on target.
Under the right circumstances, groups of people can be remarkably intelligent - smarter than the smartest person in the group individually.
Who wants to be a millionaire: call a friend (pre-determined 5 experts) gets question right 2/3s of the time, crowd gets it right 91% of the time
Google’s page algorithm would be an internet example - leveraging collective intelligence to return search results. HP started its own internal stock market to forecast which printers/products would sell. Let people buy shares at lunchtime. Google does this also - guess how many subscribers will gmail have by ______ date?
Race tracks and stock market are other examples of collective performing better than the experts.
How does this work? Aggregates judgments of people who all have some knowledge or intuition
Key is under the right conditions.
- need a way to aggregate people’s judgments, and has to be genuine bottom up decision making
- cognitive diversity - diverse crowds are far more wise than non-diverse; expands range of info crowd has access to; group has wider range of potential solutions to draw on which allows them to make better decisions (higher IQ group is less successful than a group with a lower IQ but higher level of diversity); our society is heavily invested in the idea of experts and leaders - experts and leaders are relevant but you shouldn’t rely on them solely when making judgments; experts do not generally have a good grasp of their own biases, blind spots and weaknesses - solution is to cast a wide net to solve problems; diversity also helps get around problem of group-think (devil’s advocates are very important to wise group - if someone is articulating the other position, you’ll think harder about your solution); don’t want same person to always be devil’s advocate though, it needs to rotate and arise from people’s genuine opinions
- independence - you want people to make judgments based on their own knowledge, intuition and opinions; we place too much of a premium on consensus; it leads to watered down decisions; want people to think for themselves and tap into the individuality or your group won’t be wise; independence is hard because 1. humans are by nature imitative; this is because imitation works most of the time; imitation if problematic in a group because it prevents tapping into individual knowledge; 2. most of us are concerned about our reputations; we want to appear credible so we do what our peers are doing; it is better to fail conventionally than to succeed unconventionally; being unconventional is hard
The net is a perfect medium for satisfying the “right conditions”. The knowledge we’re looking for or the info we want is often in places we would not think to look. We also overestimate our ability to figure out who experts are. The reality is that we don’t.
The net is also a tool that breaks down independence. It makes it easy to get locked into relatively small worlds (good for working on projects together or creating a small community). From the perspective of collective intelligence, this is a problem. You pay attention only to those things that are familiar. The randomized sources of info and voices is important if you want to tap into the wisdom. Keep your ties weak rather than strong.
Story of USS Scorpion - sub lost at sea but spot Navy group came up with based on scenarios they thought scorpion might have run in to.
Q&A
Psychological focus group research: They are asking people in their research what they think and what they think other people would say. This can be good but it can also be dangerous. You have to be able to filter appropriately. Within focus groups, group leaders have to be very careful not to dictate what the direction of the decision should be and not talk too much as that can effect the comments/decisions of the focus group.
Does wisdom of crowds apply to creative endeavors? No - wisdom of crowds needs a right answer in some sense and people need to agree on the problem that they are trying to solve; these two are difficult with art; its difficult to say that there is a right answer to what makes one piece of art better than another piece.
Collective intelligence models could be useful in solving crimes; juries are a little more difficult because the demand for unanimity makes it difficult right from the start.
Market research could benefit from collective intelligence - don’t ask individual what “Would you buy this?” buy but ask them “Do you think other people would buy this?” Use it like Hollywood market forecasters predict who will win an Oscar
Wrapped up with thought that we live in a strange time. We are seeing the possibilities of collective intelligence and emergence (some ideas behind Web 2.0 are attempt to harness them) At the same time, we seem to need and trust “leaders” and experts This collective intelligence model that allows info to filter from the bottom up makes us suspicious of that info (diffuses its power as info) because we don’t trust info that doesn’t come from experts Not sure what solution is or if we’ll even be able to solve it but it’s an interesting problem to work on and one with lots of potential to change society
