ArtsAutosBooksBusinessEducationEntertainmentFamilyFashionFoodGamesGenderHealthHolidaysHomeHubPagesPersonal FinancePetsPoliticsReligionSportsTechnologyTravel
  • »
  • Books, Literature, and Writing

How to Tell a Human from an AI Robot?

Updated on June 14, 2017
jackclee lm profile image

Jack is currently a volunteer at the Westchester County Archives. Jack has worked at IBM for over 28 years on museums and libraries.

Introduction

It is clear that our technology is moving towards more automation and robotics and machine learning and decision making. Every industry will be affected one way or another. In some cases, they are obvious but others are subtle. When there is no physical interface, it is hard to tell sometimes. Here are some simple tests.

- June 2017.

Background

It was Alan Turing who came up with the concept of artificial intelligence In the 1950s. His Turing test was a way to determine if a machine can fool a person in a conversation format.

I have also devised a few simple tests. Here is my list of ideas.

1. Tell a joke. A person will recognize when a joke is being made, even when if it is not that funny. On the other hand, an AI system will try to make sense out of the words and attempt at a response or if it is confused, will ask for clarification.

2. Say something satirical. It takes a person with knowledge of current affairs to recognize satire when it is said. An AI system, will be hard pressed to understand this concept.

3. Say something jibberish. This will also detect if a system is based on logic or human. A human, even a child, will know right away when something odd is said. A machine will first attempt to make sense out of a jibberish statement. It will act confused but may not recognize it as being a test.

4. Ask it to recite the number pi to the 10th digit? Most humans will not know the answer but a computer will be able to do it correctly without any delay.

5. Ask what is your favorite food? A person will say lobster or steak but an AI will not know the answer unless it was programmed into it. The next step is to follow up with what properties do you like about that food? A person will respond with the flavor, smell, texture... while a machine will say???

6. Describe one of your peak experiences. A person will respond with visiting the Grand Canyon or skydiving... while a machine will say - define peak?

7. Describe your favorite pet? A person will say my dog Sammy was the cutest thing...or I hate pets. A machine, does not have or own a pet.

8. Show it an optical illusion and ask it to describe what it sees.

Rotating Balls???

Disappearing Dots - How many black dots do you see?

Optical Illusion Sample

In the image above, I see 2 dots moving from various locations depending on where I focus my eyes. There are really 12 dots according to the creator. A machine vision system will respond with the correct answer of 12 but a human will only see 1 or 2 at any given time.

Summary

I think you get the idea. A human is so much more complex than a smart computer. As good as a computer is with facts and figures, it is no match with our human intuition and intellect. We are just smarter and street wise and we can adapt to many situations. While scientist and engineers can make a robot live like with artificial skin and expressions and speech patterns and even gestures...they cannot be human. The instantaneous response cannot be duplicated. That is what make us unique.

© 2017 Jack Lee

Comments

    0 of 8192 characters used
    Post Comment

    No comments yet.