Life on the Screen: Identity in the Age of the Internet

Chapter 5

The material below represents notes compiled by Robert Keel and Takako Nomi in their reading of Turkle's, Life on the Screen: Identity in the Age of the Internet, Simon and Schuster, 1995.. They are intended for classroom use.

The Quality of Emergence

Themes of this chapter

• Disavowal and appropriation seen in human responses to "thinking machines" A pattern of disavowal and appropriation in response to a major change in the philosophy of artificial intelligence
• Historical change in views toward Artificial Intelligence along with technological development
• From late 1960's to the mid- 1980's : Computer intelligence as being made up of a complex set of rules programmed in advance-- information processing AI
• From late 1980's : Quality of emergence in computer intelligence
• Central issues surrounding Psychology, Philosophy and Computer Science: Brain-Mind-Intelligence

Psychology

In the late 50's-- Behaviorism Study of mind in terms of stimulus and response. Exclusion of discussion of internal mental status. Study of the behavior of remembering rather than study of memory.

By the end of 60's -- Demise of Behaviorism Memory and internal states of mind. Inner status of mind in terms of logic and rules -- cognitive science -- so called Modernist intellectual aesthetic of the culture of calculation. Computers came into play -- Important metaphor to understand process taking place inside of human mind, allowing us to see internal status of mind "Thought", "Intentions", and "Goals" -- common language for both machines and human.

Views on AI from the 60's to the mid-80's

Two distinctive views of AI researchers

1. AI -- formal and logical rules that computers can follow (dominant field at that time) Systems built out of the rules. "The laws of thought" Brain as well as AI -- a single species of information process Understanding of AI through human intelligence -- Views of human intelligence as information processing
2. AI -- possibility of emergence through simultaneous interactions of independent agents withing a computer. Its underlying mathematical structure would allow them to learn from experiences

Ambivalence

Difference in Human and Machine Intelligence Views of the self -- "technology self" vs "feeling self" Human -- beyond cognitions or beyond information -- spontaneity, feelings, intentionality and sensuality

Responses in the Philosophical community in the 60's

Human intelligence

• Not reducible to propositions or rules that could be specified in advance. Intelligence arises through having a body and experiencing a changing world.
• Learning from experiences

Machine intelligence

• No such a embodiment of knowledge.
• Information processing AI -- No learning. No intuition
• However, this distinction became unclear in the 80's

In the 80's -- Questions not whether what computers do, but whether they could really be said to "understand"

Changes in the views of mind and intelligence Realms of mind -- no longer easily grasped by information processing or expert-system formalisms. "Agents" as a metaphor, which constitute intelligence. Intelligence -- distributed within a system, does not exist in any particular agent, nor reside in a isolated thinking subject, but existed with in the system as a whole Understanding -- no one part understands but the brain as a whole does

"Intelligence resides in the interaction of multiple fragments of mind, figuratively speaking, in a society of mind."

Our perceptions and the ways we understand the brain, mind, and intelligence of human, influenced our views of AI, which contributed to the development of machine intelligence. The technological development, in turn, influenced the ways we look at and understand the brain, mind and intelligence.

Emergent AI

Connectionist analogy

Connectionist -- the group of emergent AI researchers who challenged the information processing approach in the mid-1980's

Information processing AI -- Information is programmed and stored in specific locations Connectionists -- Information is not stored anywhere in particular, but rather inherent everywhere. The system's information would be evoked rather than found.

Connectionism

In order to built intelligent systems, the best way is to simulate the natural process of the brain as closely as possible -- artificial neurons and neural nets

Computers as though evolving biological organism. A system modeled after the brain -- guided not by top-down procedures, but by making connections from the bottom up, as the brain's neurons are thought to do -- the simulation of the brain. The system could learn by a large number of different connections -- unpredictable and nondeterministic. Feasibility of parallel-processing computers in the mid-1980's -- possibility to simulate parallel-processing computers on powerful serial ones. Although this resulted rather in virtual than in real, it was real enough to legitimate connectionest theories of mind Parallel computation as a new sustaining myth for cognitive science Study of emergent, parallel phenomena in the science of mind -- Cognitive psychology, Neurobiology and Connectionism

People's reactions

• People in the popular culture -- Romantic reactions: human uniqueness in constrast to computers
• People in the research community (Proponents of emergent AI) -- linking computers to the world of humans through biological and social metaphors.
  1. Computers are "notprogramed", "nondeterministic" and "spontaneous." During 70's and early 80's -- Romantic reactions.
  2. "The analytical engine has no pretensions whatever to originate anything. It can do whatever we know how to order it to perform." by Lovelace
  3. Emergent AI -- anti-Lovelace -- Continuity between computer and people.

Optimacy -- mystery and unpredictably inside its machines.

Change in approach in the field of Science "Constructing intelligence without first understanding it" Impossibility of reductionist explanation of thought "Since human intelligence was more than a set of rules, the computers that modeled it should not be about rules either"

"the story of romantic reactions to the computer presence was no longer simply about people responding to their reflection in the mirror of the machine. Now computer designers were explicitly trying to mirror of the brain. There had been a passage through the looking glass."

Mind

By late 80's

A question: Whether "the only intelligent behavior is to mirror the world with a formal theory of mind...?" The idea of neural networks may show that we intellectually behave in the world without having a theory of the world"

In late 70's and early 80's One of AI founders Minsky Mervin, information processing orientation -- Mind as a machine, mechanistic and deterministic The idea of the computer as a model of mind -- received strong criticisms because of the fear of viewing ourselves as cold mechanism (Romantic reaction)

Later, a changed image of what machines could be -- led by connectioism "The mind may be a machine, but it's not just any old machine" Connectionism -- scientific but not deterministic

Minsky The Society of Mind in 1985 -- an emergent system, "Complexity of behavior, emotion, and thought emerge from the interplay of their opposing views from their interactions and negotiations" Inner agents are still programmed in a greater degree, but their intelligence does not follow from programmed rules but emerges from the associations and connections of objects within a system

Marvin Minsky and his publications

A new response led by a connectionism: " I may be a bundle of neuron-like agents in communication" -- a view through emergent computer model

Decentered Psychology

"Emergent AI depends on the way local interactions among decentralized components can lead to overall patterns"

From centralized to decentralized: The parallel between the historical development of psychoanalysis and that of artificial intelligence

Psychoanalysis

Freudian psychology -- The idea of drive as a centralized demand that generated by the body and that provides energy and goals for all mental activity Development of the superego by internalizing important people (formation of "inner objects") compelled by our instincts.

Other Freudian theorist -- Mind as built up inner objects rather than notion of drive

Mind -- a society of inner agents "unconscious suborganization of the ego capable of generating meaning and experience, i.e. capable of thought, feeling, and perception" Within mind -- independent agencies that think, wish, and generate meaning in interaction with one another. The self -- emergence from negotiations and interactions of independent agencies

French school of psychoanalytic theory -- The idea of a centralized ego as an illusion, "only the sense of an ego emerges from chains of linguistic associations that reach no endpoint." There is no core self

Decentralized mindset -- Decentralized / Decentered views of the self

Decentralized views of mind: Ego in terms of the brain Ego not as a central authority but as an emergent system and distributed systems in the brain. Consciousness as "a technical device by which the brain represents its own workings to itself"

Connectionist analogy -- The question is rather how we imagine the machine handles information (We really don't understand connectionism without mathematical sophistication )

Connectionism as a metaphor rather than technical understanding of theory -- What the machines do is less important than how people think about them

the Ear of Post Modernism: Feeling of fragmentation

Appropriation of images and metaphors rather than logic and reason Connectionist ideas of AI offer almost tangible objects-to-think-with when we approach mind The brain and a machine, Machines -- "Neuron-like entities" The ideas of decentralized mind and the self, denying the ego as the central executive of the mind

Use of imagination

1. Emergent theories of mind
2. Recentralization theories -- centralized authority behind the idea of emergence Associations of neural pathways as "Executive role" -- an ability to recognize patterns

What are the boundaries between machines and human?

The first boundary displacement : the displacement of Romantic reactions -- obstructed by psychophramacology, suggesting that the process that underlies human motions are fairly "mechanical", predictable,and controllable.

The second boundary displacement: Shift in people's sense of differences from computers -- shift away from the domain of intelligence to the domain of biological life

Computers are accepted as intelligent, but people are special because we are alive

Now, the issue is LIFE