Beyond the Turing Test
The Turing Test, developed by British scientist Alan Turing and now familiar to viewers of the 2014 movie “The Imitation Game,” has long been considered the gold standard for the measurement of human intelligence. If, by its responses to a set of challenging verbal questions, a computer can fool a careful observer, then the computer would be deemed intelligent.
Now it is being increasingly recognized that no single set of questions, delivered in a single format, can determine whether a machine is intelligent. Rather, as described in an article from Science, a new and improved Turing Test must incorporate several measures due to the expanding capabilities of artificial intelligence. To quote from the article, in a new Turing Championship that would include a greater number of benchmarks and questions for computers, “the proposed challenges acknowledge that intelligence has multiple dimensions, from language acquisition to social awareness, that are best tackled piece by piece.”
Thus, a modern Turing Test should seek to measure various capabilities present in the human mind as a determinant of whether a respondent is human. And indeed, the list of capacities in this article—from physical movement to the ability to collaborate—is quite reminiscent of the ensemble of multiple intelligences.