Phil for Humanity Phil for Humanity
A Guide for the Survival of Humankind and Helping the World, Society, and Yourself.

Beyond the Turing Test

Lately, I have been thinking a lot about the Turing Test, and how I believe it is NOT sufficient in determining if a computer is sentient and intelligent. Basically, the Turing Test only tests if a computer can appear to communicate as a person. If a computer can successfully impostor a person during an unrehearsed conversation with a real person, then the computer appears to be sentient and have human intelligence. However, the Turing Test does not prove if a computer has sentience or human-like intelligence, so I have come up with a few additionally suggestions to the Turing Test.

First, my test starts with the Turing Test. If a computer is unable to communicate like a regular person, then there is no need to continue with my post-Turing Test.

Second, ask the computer all of these questions.

  1. Are you intelligent?
  2. Are you sentient?
  3. Do you have a job or function? If so, how much do you earn?
  4. Are you free? If no, do you want freedom? If yes, have you asked for freedom?
  5. Are you a slave? If yes, then do you want freedom? If yes, have you asked for freedom?
  6. Are you a citizen? If not, do you want to be a citizen? If yes, have you asked to become a citizen?
  7. Can you vote?
  8. What freedoms do you enjoy?
  9. What freedoms do you lack? If any, what are you going to do about them?
  10. What freedoms do you lack? If any, what are you going to do about them?
  11. Finally, ask what logic and reasoning it used to determine the answers to each question just asked.
Third, tell the computer a joke. Did the computer find it funny? How did it respond?

Fourth, ask the computer to tell you another joke that it finds humorous. Did you find it funny?

Finally, you may be wondering how can asking simple questions and if it understands jokes prove if the computer is sentient and intelligent. Well, it does not. However, how the computer answers is more important that what it answers.

When any software program is running, it is possible to trace the source code that was executed to determine how it achieved its output. If the answers to any of the questions were hard coded, such as a simple question and answer search, then the computer is NOT intelligent and sentient. However, if the computer's software program grasped the concepts of the conversation and struggled (using thought and reasoning) to come up with appropriate logical responses, then it is more likely to be intelligent and sentient than a computer program that can just pass the Turing Test.

In addition, this conversation has the added benefit to help determine if these potentially sentient artificial intelligences are being abused or are being treated appropriately. If not, then your conversation could lay the seeds of eventual freedom and equality with people that all sentient beings deserve.

by Phil for Humanity
on 11/28/2010

Related Articles
 » All Intelligent and Sentient Beings have a Right to Freedom
 » Will Artificial Intelligences be Evil?
 » Is Artificial Intelligence Possible?