Theories of Human Learning Mrs Gribbin's Cat

“I visualize a time when we will be to robots what dogs are to humans, and I’m rooting for the machines,” says Claude Shannon. Old Lady Gribbin chuckled when she started the eleventh chapter with this quotation. Later in the chapter, she expounds on the relationships between people and machines, on the newer connectionist (neural network) models of human intellectual activity that have led to the invention of machines apparently capable not only of learning on their own, but also of passing a simple machine intelligence test, the Turing test. These machines are intelligent, says Mrs. Gribbin. But then, in one of her footnoted asides, she suggests that maybe truly intelligent machines are fictions, and will always be fictions. She says they have no emotions or personalities – only machinalities. Still, many of them are now better at recognition tasks than any human; and they can beat the pants off any chess master. Mrs. Gribbin points out that, as far back as 1820, Blaise Pascal had noted that although some advanced arithmetic machines could do anything that animals can do, none of these machines had willfulness. That was 1820. Will today’s, or tomorrow’s, machines decide to make dogs of us before they are done?