Class 40: Evaluating AI

Held Tuesday, May 2, 2000

Overview

Today we concluder our discussion of AI by revisiting the criteria by which we might judge a program ``intelligent''.

Question 40 for discussion today: Do you consider the Turing test a reasonable way to evaluate the the success of an AI program? Why or why not?

Question 41 to prepare for Monday: Summarize the most important thing you've learned from the reading.

Notes

• Next week, we'll move on to some topics in the theory of computation. Essentially, our goal is to talk about what can and cannot be computed, but we need some background first.
• The following week, we'll try to move on to topics in e-commerce.
• For Monday, read Dewdney 2 (Finite Automata), 7 (The Chomsky Hierarchy), and 66 (Church's Thesis)

Contents

• Administrative
• Some Starting Questions

Summary

• The Turing Test
• Is it enough?
• If not, what is?

Administrative

• Discussion leader: Sam
• Recorder: Ivy
• Devil's advocate: Kevin

Some Starting Questions

• What is the Turing test?
• Is it an appropriate test of human intelligence? Why or why not?
• If it is not appropriate an appropriate test, what would be?