Cass Sunstein, Infotopia: How Many Minds Produce Knowledge (2006)
Author Background:
American legal scholar; 27 years at University of Chicago Law School and White House Administrator of Office of Information and Regulatory Affairs
Thesis: 57 – It cannot be shown that deliberating groups generally arrive at the truth. … Much of the time, deliberating groups do quite poorly at aggregating the information that their members have. …When individuals show a strong bias or a clear tendency toward error, deliberating groups often show a greater bias and hence a greater tendency towards error.
179 – The central point is that it is possible to give numerous minds access to certain processes, and that in some circumstances, at least, this access is an engine for rapid improvements.
xiii – Even the most informed people have direct knowledge of only a tiny fraction of the “facts” on which their lives depend.
– Four different methods of eliciting and aggregating information … statistical average of the independent judgment of their members … deliberation .. price system … the Internet. They all carry some baggage that will effect the outcome, the question is really which one is bets and produces the most reliable result.
9 –information cocoons: communications universes in which we hear only what we choose and only what comforts and pleases us. … decisions will not be adequately challenged from the inside … preconceptions will become entrenched (particularly bad in politics)
12 – Countless groups do badly not in spite of deliberation but because of it. The problem is that deliberating groups often do not obtain the knowledge that their members actually have.
13 – informational influences which cause group members to fail to disclose what they know out of respect for the information publicly announced by others.
14 –social pressures which lead people to silence themselves to avoid the disapproval of peers and supervisors
14-6 profit-driven decision making instruments, like prediction markets, are better than deliberating groups but not flawless
15 – Markets can incorporate falsehood as well as truth … prediction markets, a new innovation, often do startlingly well simply because they are so effective at pooling information.
16 – problems posed by informational pressure and social influences apply in all domains. They infect our most fundamental judgments about morality and policy, not merely … facts.
32 – If a significant number of people are more likely to choose the correct answer than any of the incorrect ones, and if errors are randomly distributed, then the average judgment will be quite reliable.
34 – people are susceptible to anchors, in the form of starting points that can greatly bias their judgment
51 – a constitution that ensures the “jarring of parties” and “differences of opinion” would “promote deliberation”
52 deliberation is humanity’s favored way of coming to a decision, but sometimes group deliberation doesn’t work
60 – brainstorming is actually most beneficial when carried out initially in private, the interacting group then being used as a forum for combining and evaluating these individually produced ideas.
64-5 in groups, the majority is decidedly influential even when wrong
75 – deliberating groups typically suffer from four problems. They amplify the errors of their members. They do not elicit the information that their members have. They are subject to cascade effects … show a tendency to group polarization.
77 – Most people are also strikingly vulnerable to framing effects, making different decisions depending on the wording of the problem.
79 –Groups are more likely than individuals to escalate their commitment to a course of action that is failing—and all the more so if members identify strongly with the groups of which they are a part.
93 – extremism and even fanaticism and terrorism. All these can be fueled, not reduced, by deliberation.
95 – As people gain confidence, they usually become more extreme in their beliefs.
95 corroboration gives a boost to opinion that may initially be weakly held
105 predictive markets provide a monetary incentive for sharing private knowledge, especially if the contribution is anonymous
106 prediction markets eliminate individual errors, not magnify them
123 – Hayek’s argument is that human morality is itself the product of many minds, making their decisions over long periods of time, in a way that produces a set of principles that no individual mind, and no theory, is likely to be able to capture.
124 – For morality, a particular problem is that social pressures often force people to silence themselves, leaving hidden profiles.
139 a way that prediction markets counter bias is that unbiased traders are able to profit off of biases, which tends to correct the “price” – like a moving line on a sporting event
143 prediction markets can’t judge values or morals, just facts
147 –collaborative filtering, the process of figuring out what you’re likely to like by investigating the tastes of minds that are like yours.
154 – Wikipedia works because those who know the truth, or something close to it, are usually more numerous and more committed than those who believe in a falsehood. … In general, the worse the error, the faster it will be noticed and fixed. … In some areas, what is true is greatly disputed, and it is hard to find an impartial arbiter.
168 – open source software provides a method by which decentralized bits of private knowledge and creativity can be elicited and used
201 – If people are asked to think critically rather than simply join the group, and they are told that the group seeks and needs individual contributions, then disclosure is more likely.
207 FDR privately said he agreed with them to people holding opposing positions – to make them give their best effort when debating in more public venues
207 – leaders and high-status members can do groups a great service by indicating their willingness and even desire to hear information that is held by one or a few members and that might otherwise receive little or no attention
211 – Many experimenters have found that protection of genuine dissenting views can enhance group performance.
223 “a strong norm in favor of critical thinking can reduce some of the most damaging pressures, and hence ensure that people will hear from many minds rather than a few”
Implications for Strategy:
- The basic goal should be to increase the likelihood that deliberation will do what it is supposed to do: elicit information, promote creativity, improve decisions.
- Deliberations can be helpful to solve hard or confusing problems; but they can also cement extreme views
- The internet is not convincing people of alternative views, rather it is making views more extreme
- Current military attitude is toward more aggregation of every piece of data, with little apparent consideration for the limitations in the PROCESS of aggregation: each process has its strengths and weaknesses