IEEE Future Directions
Related Articles
As we were revising the second Symbiotic Autonomous Systems White Paper the authors were in disagreement on the Turing test having been passed. Now, the very fact that such a discussion is taking place points out 2 things:
- computers (software) have progressed to the point that it may be debatable if it can be mistaken by a human
- as we get closer to the human-ness the boundary gets more and more blurred and it gets difficult to be defined.
Indeed we have seen several instances of programs that have been able to fool a human interviewer in a given subject area. These areas have been expanded and we have seen the questioning moving into arts, poetry… and still the computers are proving to be able to pass the test.
However, as we have got closer to a human interaction we have also started to wonder if having a human like (indistinguishable) interaction can really prove human-ness. According to John Searle, a philosopher who wrote a paper on “Minds, Brains and Programs” having a human like interaction does not prove that the interacting entity is thinking, hence it does not prove it is human like. He proposed a though experiment, the Chinese Room, in which he would manipulate Chinese characters with the help of a computer and would be able to associate to strings of Chinese characters received by a Chinese person through a hole in the wall other sets that would make sense and pass them through the hole to the person on the other side. For that person the feeling would be to “interact” with a Chinese person whilst in reality he would be interacting with an automaton able to associate input string to output strings in a convincing way but nevertheless having no inkling of what it means.
I should confess that in my view the Turing test has been passed, although I accept that others may not agree. I am taking a pragmatic view: if I cannot distinguish a computer from a person through an interaction, well, in that particular interaction I would consider them equivalent. I am a fan the:
If it looks like a duck, swims like a duck, and quacks like a duck, then it probably is a duck
I know it is wrong, but to any practical effect it is right! Apparently this is also the point of view of the state of California where Jerry Brown, the governor, has signed a law that will go into effect on July 1st 2019, imposing companies using bots (chat-bots) to make it clear at the beginning that you are not talking to a person but to a machine.
I feel this is important and it is marking a sort of technological – cultural divide. We have reached a point where we can no longer be aware if the other party is a human or a machine. We need somebody to tell us. And it is a slippery slope. In the coming decades as humans will become augmented through technology the line between humans and cyborg will get fuzzy and we will probably need a law demanding our fellow citizen to let us know if he is a cyborg…
If you are on the side of people feeling that computers cannot be mistaken for a real person take a minute to look, actually, listen to Google Duplex making a hairdresser appointment…
