“AI is one of the most important things humanity is working on,” said Google CEO Sundar Pichai earlier this year. “It is more profound than, I dunno, electricity or fire.”
But day-to-day developments in the field are rather mundane, according to Gary Marcus, a professor of psychology and neural science, and Ernest Davis, a professor of computer science, both at New York University. They cite Pichai earlier this month demonstrating Google Duplex making a phone call to schedule a hair-salon appointment—a limited accomplishment in a closed domain, something an American with a phrase book could do in France or Germany.
“Schedule hair-salon appointments?” Marcus and Davis write in The New York Times. “The dream of artificial intelligence was supposed to be grander than this—to help revolutionize medicine, say, or to produce trustworthy robot helpers for the home.”
The problem with current implementations, they say, lies with the machine-learning techniques employed today. They advocate a return to knowledge engineering, whose goal is “…not to create programs that would detect statistical patterns in huge data sets but to formalize, in a system of rules, the fundamental elements of human understanding, so that those rules could be applied in computer programs. Rather than merely imitating the results of our thinking, machines would actually share some of our core cognitive abilities.”
Today’s strategy has not worked out, they write, concluding, “If machine learning and big data can’t get us any further than a restaurant reservation, even in the hands of the world’s most capable AI company, it is time to reconsider that strategy.”