1
Social Systems Designing Digital Systems that
Support Social Intelligence

• Thomas Erikson
• AI and Society
• Presented by Rosta Farzan
• PAWS Group Meeting
• April 27, 2007

2
Goal
Examines social intelligence in face-to-face
behavior

• Supporting social intelligence in online systems

Social Proxy A Minimalist visualization of
presence and activities of participants in online
systems
3
Outline

• Social intelligence
• Social intelligence in the face-to-face world
• Design of social intelligence in online systems
• Social Proxies
• Examples of social proxies
• User experience
• Design principles
• Discussion

4
Social Intelligence

• The ways in which groups of people manage to
  produce coherent behavior directed towards
  individual or collective ends

• Why do we care about social intelligence?
• Groups may produce qualitatively better and
  quicker results than individuals
• Solution produced collectively is more preferred

5
Social Intelligence in Face-to-Face World

• A tale of two doors
• A door opening to a busy hallway
• Solutions
• A sign which says open door slowly
• Inserting a window in the door
• Visibility and Accountability
• Street crossing

6
Social Intelligence in Online Systems

• Geographically separated people
• Design
• Invoking social norms
• Making physical and behavioral cues visible in
  online environments
• Possibilities
• 3D virtual environment with avatars
• Using videos

Significant amount of bandwidth, screen space Not
scalable Associated with games Privacy issues
7
Social Proxy

• Minimalist graphical representation of socially
  salient aspects of an online interaction

8
Examples of Social Proxies
Social Proxy for chat environment
Circles represent chat room Dots represent
users Active users are close to circles hub
Appropriate for synchronous patterns of
interaction
Timeline represents users activity Users leave
flat traces when logged on Traces are blipped
when users talk
Appropriate for patterns of activity over time
9
More Examples

• Lecture proxy

Dots represent people Dots are positioned
according to how much each person has spoken
during the last 3 minutes
Lecture Norm No audience interruption
An audience member interrupting
Many interruption
10
Large Scale Example

• Task proxy for organization-wide activities
• Facilitate performance of tasks that require the
  coordination of many individuals

11
User Experience

• Chat proxies
• Engaging and informative
• Users making inferences not necessarily correct
• Part of life as social beings
• Easy to learn
• Ice-breakers
• Topic for conversation

12
Design Principles

• Everyone sees the same thing no user
  customization
• Portray actions, not interpretation
• Should allow deception
• Support micro/macro readings
• Ambiguity is useful suggest rather than inform
• Use a third person point of view

13
Discussion