Artificial Intelligence can be a subjective term. However, researchers have now quantified just how intelligent machines could be.
When scientists gave ConceptNet, a smart machine, a verbal IQ test designed for small children, the results were telling.
The machine's test score matched intelligence that was average for a four-year-old child, but below average for five- to seven-year-olds.
It's an impressive score given that we are talking about machines.
However, a key factor to note is the machine's inability to comprehend questions in the way a child would.
When a team from the University Of Illinois asked the computer questions such as “where can you find a teacher?” it based its answer on concepts.
In this case, it understood the request as three different concepts: "find," "teacher" and "find teacher."
When processing all three requests, it was able to come up with a sensible answer.
However, when lead researchers forced it to only use one concept "find teacher" the answers were completely random: "band" and "piano" -- a marked difference from how human beings process information.
The research was carried out by Stellan Ohlsson. Commenting on the results he said: "“ConceptNet does well on Vocabulary and Similarities, middling on Information, and poorly on Word Reasoning and Comprehension.”
MIT Technology Review explained that Ohlsson's results pose a more urgent question: "what kind of AI machine might we be grappling with in 2018?"
It explains that while it has taken 60 years to produce a machine with reasoning capabilities similar to a human child, "exponential improvements" over the next six years could see "dramatic improvement."
Maybe our robotic overlords will grow up and hit their teens.