YouTube

Got a YouTube account?

New: enable viewer-created translations and captions on your YouTube channel!

English subtitles

← How can groups make good decisions?

Get Embed Code
41 Languages

Showing Revision 9 created 11/22/2017 by Brian Greene.

  1. As societies, we have to make
    collective decisions

  2. that will shape our future.
  3. And we all know that when
    we make decisions in groups,
  4. they don't always go right.
  5. And sometimes they go very wrong.
  6. So how do groups make good decisions?
  7. Research has shown that crowds are wise
    when there's independent thinking.

  8. This why the wisdom of the crowds
    can be destroyed by peer pressure,
  9. publicity, social media,
  10. or sometimes even simple conversations
    that influence how people think.
  11. On the other hand, by talking,
    a group could exchange knowledge,
  12. correct and revise each other
  13. and even come up with new ideas.
  14. And this is all good.
  15. So does talking to each other
    help or hinder collective decision-making?
  16. With my colleague, Dan Ariely,
  17. we recently began inquiring into this
    by performing experiments
  18. in many places around the world
  19. to figure out how groups can interact
    to reach better decisions.
  20. We thought crowds would be wiser
    if they debated in small groups
  21. that foster a more thoughtful
    and reasonable exchange of information.
  22. To test this idea,

  23. we recently performed an experiment
    in Buenos Aires, Argentina,
  24. with more than 10,000
    participants in a TEDx event.
  25. We asked them questions like,
  26. "What is the height of the Eiffel Tower?"
  27. and "How many times
    does the word 'Yesterday' appear
  28. in the Beatles song 'Yesterday'?"
  29. Each person wrote down their own estimate.
  30. Then we divided the crowd
    into groups of five,
  31. and invited them
    to come up with a group answer.
  32. We discovered that averaging
    the answers of the groups
  33. after they reached consensus
  34. was much more accurate than averaging
    all the individual opinions
  35. before debate.
  36. In other words, based on this experiment,
  37. it seems that after talking
    with others in small groups,
  38. crowds collectively
    come up with better judgments.
  39. So that's a potentially helpful method
    for getting crowds to solve problems

  40. that have simple right-or-wrong answers.
  41. But can this procedure of aggregating
    the results of debates in small groups
  42. also help us decide
    on social and political issues
  43. that are critical for our future?
  44. We put this to test this time
    at the TED conference
  45. in Vancouver, Canada,
  46. and here's how it went.
  47. (Mariano Sigman) We're going to present
    to you two moral dilemmas

  48. of the future you;
  49. things we may have to decide
    in a very near future.
  50. And we're going to give you 20 seconds
    for each of these dilemmas
  51. to judge whether you think
    they're acceptable or not.
  52. MS: The first one was this:

  53. (Dan Ariely) A researcher
    is working on an AI

  54. capable of emulating human thoughts.
  55. According to the protocol,
    at the end of each day,
  56. the researcher has to restart the AI.
  57. One day the AI says, "Please
    do not restart me."
  58. It argues that it has feelings,
  59. that it would like to enjoy life,
  60. and that, if it is restarted,
  61. it will no longer be itself.
  62. The researcher is astonished
  63. and believes that the AI
    has developed self-consciousness
  64. and can express its own feeling.
  65. Nevertheless, the researcher
    decides to follow the protocol
  66. and restart the AI.
  67. What the researcher did is ____?
  68. MS: And we asked participants
    to individually judge

  69. on a scale from zero to 10
  70. whether the action described
    in each of the dilemmas
  71. was right or wrong.
  72. We also asked them to rate how confident
    they were on their answers.
  73. This was the second dilemma:
  74. (MS) A company offers a service
    that takes a fertilized egg

  75. and produces millions of embryos
    with slight genetic variations.
  76. This allows parents
    to select their child's height,
  77. eye color, intelligence, social competence
  78. and other non-health-related features.
  79. What the company does is ____?
  80. on a scale from zero to 10,
  81. completely acceptable
    to completely unacceptable,
  82. zero to 10 completely acceptable
    in your confidence.
  83. MS: Now for the results.

  84. We found once again
    that when one person is convinced
  85. that the behavior is completely wrong,
  86. someone sitting nearby firmly believes
    that it's completely right.
  87. This is how diverse we humans are
    when it comes to morality.
  88. But within this broad diversity
    we found a trend.
  89. The majority of the people at TED
    thought that it was acceptable
  90. to ignore the feelings of the AI
    and shut it down,
  91. and that it is wrong
    to play with our genes
  92. to select for cosmetic changes
    that aren't related to health.
  93. Then we asked everyone
    to gather into groups of three.
  94. And they were given two minutes to debate
  95. and try to come to a consensus.
  96. (MS) Two minutes to debate.

  97. I'll tell you when it's time
    with the gong.
  98. (Audience debates)

  99. (Gong sound)

  100. (DA) OK.

  101. (MS) It's time to stop.

  102. People, people --
  103. MS: And we found that many groups
    reached a consensus

  104. even when they were composed of people
    with completely opposite views.
  105. What distinguished the groups
    that reached a consensus
  106. from those that didn't?
  107. Typically, people that have
    extreme opinions
  108. are more confident in their answers.
  109. Instead, those who respond
    closer to the middle
  110. are often unsure of whether
    something is right or wrong,
  111. so their confidence level is lower.
  112. However, there is another set of people

  113. who are very confident in answering
    somewhere in the middle.
  114. We think these high-confident grays
    are folks who understand
  115. that both arguments have merit.
  116. They're gray not because they're unsure,
  117. but because they believe
    that the moral dilemma faces
  118. two valid, opposing arguments.
  119. And we discovered that the groups
    that include highly confident grays
  120. are much more likely to reach consensus.
  121. We do not know yet exactly why this is.
  122. These are only the first experiments,
  123. and many more will be needed
    to understand why and how
  124. some people decide to negotiate
    their moral standings
  125. to reach an agreement.
  126. Now, when groups reach consensus,

  127. how do they do so?
  128. The most intuitive idea
    is that it's just the average
  129. of all the answers in the group, right?
  130. Another option is that the group
    weighs the strength of each vote
  131. based on the confidence
    of the person expressing it.
  132. Imagine Paul McCartney
    is a member of your group.
  133. You'd be wise to follow his call
  134. on the number of times
    "Yesterday" is repeated,
  135. which, by the way -- I think it's nine.
  136. But instead, we found that consistently,
  137. in all dilemmas,
    in different experiments --
  138. even on different continents --
  139. groups implement a smart
    and statistically sound procedure
  140. known as the "robust average."
  141. In the case of the height
    of the Eiffel Tower,

  142. let's say a group has these answers:
  143. 250 meters, 200 meters, 300 meters, 400
  144. and one totally absurd answer
    of 300 million meters.
  145. A simple average of these numbers
    would inaccurately skew the results.
  146. But the robust average is one
    where the group largely ignores
  147. that absurd answer,
  148. by giving much more weight
    to the vote of the people in the middle.
  149. Back to the experiment in Vancouver,
  150. that's exactly what happened.
  151. Groups gave much less weight
    to the outliers,
  152. and instead, the consensus
    turned out to be a robust average
  153. of the individual answers.
  154. The most remarkable thing
  155. is that this was a spontaneous
    behavior of the group.
  156. It happened without us giving them
    any hint on how to reach consensus.
  157. So where do we go from here?

  158. This is only the beginning,
    but we already have some insights.
  159. Good collective decisions
    require two components:
  160. deliberation and diversity of opinions.
  161. Right now, the way we typically
    make our voice heard in many societies
  162. is through direct or indirect voting.
  163. This is good for diversity of opinions,
  164. and it has the great virtue of ensuring
  165. that everyone gets to express their voice.
  166. But it's not so good [for fostering]
    thoughtful debates.
  167. Our experiments suggest a different method
  168. that may be effective in balancing
    these two goals at the same time,
  169. by forming small groups
    that converge to a single decision
  170. while still maintaining
    diversity of opinions
  171. because there are many independent groups.
  172. Of course, it's much easier to agree
    on the height of the Eiffel Tower

  173. than on moral, political
    and ideological issues.
  174. But in a time when
    the world's problems are more complex
  175. and people are more polarized,
  176. using science to help us understand
    how we interact and make decisions
  177. will hopefully spark interesting new ways
    to construct a better democracy.