A live-chatting computer has passed the artificial-intelligence (AI) milestone Turing Test in London, as it fooled a third of the people into thinking they were speaking to a human. The father of modern computer science, Alan Turing, developed the Turing test 65 years ago to answer whether machines can think. To pass, a computer programme must be mistaken for a human in over 30% of its five-minute keyboard conversation. The event marked the 60th anniversary of Turing’s death, and took place nearly six months after he received a posthumous royal pardon for his 1952 homosexuality conviction.
The Turing Test is fascinating. How we judge “intelligence” has been a philosophical quandary for centuries. We judge animals to be “intelligent”, for instance—though we do not know if they possess the same self-aware thinking and feeling we associate with human consciousness. If humans are mere biological machines, and no different from an insect or dog in terms of the ability to solve problems, then perhaps if we can interact with computers as equals, we will consider them as “intelligent.”
However, a typical Reader’s Digest-type disclaimer of AI is that although a computer can answer any number of questions, it cannot ask one. It is the inquisitiveness of humans which makes us different from robots. Moreover, not only are there differences in the levels of intelligence among humans, the faculty can grow or improve with age. This is unlike ever to be true of a computer, whose acumen will remain static although other, more intelligent computers can be created. In the London event, for instance, the computer was presented to the audience as a 13-year-old boy, who will remain an adolescent till “he” or it is dismantled. In the ultimate analysis, therefore, it is the creator of computers who will always be seen as more intelligent than an answering machine.