The Humanity of Robots

The Wall Street Journal ran an interview this week with Luke Muehlhauser, the Executive Director of the Singularity Institute. The Journal asked: Will Artificial Intelligence Make us Obsolete? Muehlhauser's answer was, well yes. In his words:

Cognitive science has discovered that everything the human mind does is done by information processing and machines can do information processing too.

The first statement is clearly false, or at least depends on a strangely mixed up idea of "information processing." The old determinist canard that humans are simply complex machines has not been proven or discovered by cognitive science. And even if humans do process billions upon billions of bits of information it is not at all clear that such a humanly fallible process is reproducible. That is not the claim that cognitive science can make.

But cognitive science can claim that machines can be built that act in ways that are so like humans as to be almost nearly indistinguishable from them. Or, they can even be better than humans in doing many quintessentially human tasks. So machines can not only beat humans at chess, they can make moves that seem like moves only a human could have made, as Gary Kasparov learned to his dismay in the second game of his rematch with Deep Blue. Machines can create paintings that appear to be fully creative, as does Aaron, the painting machine created by artist and computer scientist Harold Cohen. And machines can increasingly make ethical decisions in warfare, as the robo-ethicist Ron Arkin has argued—decisions that are more humane than those made by human warriors.

Too much of the debate over artificial intelligence is caught up in the technical and really irrelevant question of whether machines can fully replicate human beings. The point is that if machines act "as if" they are human, or if they are capable of doing what humans do better than humans, we will gradually and continually allow machines to take over more and more of the basic human activities that make up our world. Already computers make most of the trades on Wall Street and computers are increasingly used in making medical diagnoses. Computers are being used to educate our children and write news stories. Caregivers for the elderly are being replaced by robotic companions. And David Levy, artificial intelligence researcher at the University of Maastricht in the Netherlands, argues that we will be marrying robots in the near future. It is not that these robotic lovers or artificial artists are human, but that they love and paint in ways that do or will soon pass the Turing test: they will be impossible to distinguish from human works.

Undoubtedly one reason machines are acting more human is that humans themselves are acting less so. As we interact more and more with machines, we begin to act predictably, repetitively, and less surprisingly. There is a convergence at foot, and it is the dehumanization of human beings as much as the humanization of robots that should be worrying us.

-RB