Ethical & Social Impact of Computing, SYCS 390

This site is used in conjunction with the course. It does not contain all the material given in class.

There may be errors in this material, and it should not be used in place of material given in the classes.

Can Machines Think?

Deep Blue beat Garry Kasparov, Chess Champion of the World in 1997.
Is Deep Blue Intelligent?

Can Deep Blue Think?

• Mass media - ?
  
• Gary Kasparov - ?
  
• Computer scientists - ?
  
• IBM - ?
  
• General public - ?
  

Did Deep Blue pass the Turing Test?

Turing Test
An interviewer, sitting in a separate room, asks a series of questions randomly directed to either a computer or a human. Based on the answers, the interviewer must distinguish computer responses from human responses. If the interviewer is unable to do so, then the computer is intelligent.

Intelligence
The ability to learn or understand from experience; ability to acquire and retain knowledge; mental ability - Webster's Dictionary, 2nd Ed., 1977.

Artificial Intelligence
Term first used by John McCarthy of Stanford University in 1956.

What is AI?
Different Perspectives

• Intelligence - making machines intelligent by making them act as we would expect people to act.
  
• Research - the study of how machines might behave like people.
  
• Business - a set of powerful tools and methodologies for using these tools to solve business problems.
  
• Programming - the study of symbolic programming, problem solving, and searching.
  

Another Perspective of AI (from Ben Schneiderman, U. Maryland, 1998.)

• Making machines intelligent by making them act as we would expect people to act is mimicry.
  
• People are not computers and computers are not people.
  
• Users want powerful tools that amplify their capabilities and enable them to do their tasks.
  

What is an expert system?
For that matter, what is a human expert?

What is an expert system?

• A computer system designed to simulate the problem-solving behavior of a human who is an expert in a narrow domain (Peter Denning, GMU)
  
• A computer program that behaves like a human expert in some useful ways (Patrick Winston, MIT)
  

Expert system architecture

• Knowledge Base - database of rules and facts
  
• Inference Engine - general problem solving knowledge or methods
  
• User Interface - language for users to express queries
  

Types of Expert Systems

• Shallow - knowledge base stores more facts than rules, designed for speed, conclusions often straight-forward consequences of information in knowledge base.
  
• Deep - have underlying model of phenomena in their domain, some conclusions not obvious.
  

Examples of Expert Systems

• Shallow - XCON (expert configurer), an ES that helped configure VAX computer systems (developed by DEC)
  
• Deep - Dendral, an ES for analyzing mass spectogram data, accepts mass spectograms and chemical formulae and finds molecular structures (Feigenbaum, et. al.)

Human Expert
Experts in a particular field can:

• Solve problems
  
• Can explain what they do
  
• Judge reliability of their conclusions
  
• Learn from experience
  
• Know when they do not know
  
• Transfer knowledge between domains
  
• Break rules
  
• Determine relevance
  
• Degrade gracefully
  

Information added in 2002, Copyright Dr. Wardle 2002

Smart Environments

• Sociable technology (has emotion):
  robotic dogs for the elderly, sociable toys for children
• Augmented Reality:
  walls that change in internal offices
• Intelligent Homes:
  Controllable & Adaptable
• Smart Medical Home:
  smart aids for the elderly & infirm