Artificial Intelligence machine has IQ of 4-year olds 

Here,what you will read next is probably going to amaze you. It's the current discovery of the the potential differences between human IQ and Machine IQ...

How 'intelligent' is current Artificial Intelligence? Machines which are hugely better at handling data or playing chess or even pattern recognition are well-known. But clearly, this doesn't give us an idea of how they stand in comparison to human intelligence.

A team of scientists seem to have partially cracked this conundrum by comparing the 'intelligence quotient' or IQ of one of the world's most powerful artificial intelligence machines with human intelligence. And their results are not surprising: computers have to go a long way before they can catch up with humans. One of the best machines could manage an IQ score of 4-year old child at the most.

Stellan Ohlsson at the University of Illinois and his colleagues gave the standard IQ test to ConceptNet 4, a powerful AI machine built at the Massachusetts Institute of Technology (MIT). To do this they had to first convert the test designed for children into a language understood by the computer.

 This test measures children's performance in five categories: information, vocabulary, word reasoning, comprehension, and similarities, according to MIT Technology review.

"ConceptNet does well on Vocabulary and Similarities, middling on Information, and poorly on Word Reasoning and Comprehension," say Ohlsson and his colleagues, according to the Technology Review.

For example, in the comprehension category, the machine was asked "Why do we shake hands?" ConceptNet 4 searched its databases for concepts of "shake", "hand" and "shake hands." And its answer was "epileptic fit."

When the researchers forced the machine to look for only single concepts ("shake" and "hands") ConceptNet came up with more satisfactory answers "thanks," "flirt," and "meet friend."

The Technology Review describes various other examples of often bizarre answers that the machine kept coming up with. But the key finding was that many of the wrong answers were so wrong that even children wouldn't give such answers.

A detailed analysis by Ohlsson and his colleagues of the answers given by ConceptNet to the verbal intelligence test (called WPPSI-III VIQ) showed that the machine scored like a four year old child would.

"The ConceptNet system scored a WPPSI-III VIQ that is average for a four-year-old child, but below average for five- to seven-year-olds," they say. 

Source:  Times of India

Image credits: techemergence

Affiliate disclosure

What's your opinion?