When we look at the picture on the left, it’s obvious that something’s wrong with the picture on the bottom. The human brain’s capability to integrate knowledge is what allows us to recognize that keyboards and computers fit well together while plants and computers… not so much.
Christof Koch, Cognitive and behavioural biology professor at the California Institute of Technology and Giulio Tononi, professor in consciousness science at
the University of Wisconsin-Madison, have explored the notion of consciousness and propose some simple tests that could be used to find out if a machine were truly conscious or not.
What would these tests consist of? The Turing test is what’s commonly currently used to test the intelligence of machines. A person must judge whether they are communicating with another human or an intelligent machine via computer. To date, machines have been able to fool people during this test but after a certain point in the conversation it becomes clear to the human judge that what they are talking to is not made of flesh and blood.
Yet how could we test if a machine were ever truly conscious? In Scientific American’s A Test for Consciousness, Koch and Tononi propose the use of different types of tests which rely on a subject’s ability to perceive enormous amounts of integrated knowledge: simple puzzles. The image-analysis systems of today’s most intelligent machines, which are capable of identifying one specific face out of a million options, would be stumped when trying to choose the correct picture of the two seen above. Likewise, the placement of a few black strips on some pictures easily interferes with many image-analysis systems and renders today’s machines unable to pick out the matching portions of the pictures.
However, as machines are built more and more according to biological principles, Koch and Tononi maintain that one day machines will easily be able to pass such tests “and when they do they will share with us the gift of consciousness- the most enigmatic feature of the universe”.