It all began in 1950, when Alan Turing published an article in whose opening paragraph he asked: can a machine think?
The reason for something so unusual at the time was linked to a proposal also put forward by Alan Turing shortly before the outbreak of World War II. The brilliant mathematician had argued in 1936 that by combining a simple mechanism with an appropriate procedure, one could achieve a “universal machine” capable of simulating any device based on the execution of algorithms. Suddenly, Alan Turing introduced the concepts of hardware —the inner workings of machinery— and software —the instructions that make it operate— while at the same time opening the door to the science of computing. The impact was considerable. But none of Turing’s ideas led to as much controversy as the question of whether a machine can think.
In particular because his answer was brilliant. By proposing a simulation game, Turing suggested that if an unseen respondent —answering in writing, of course— replies to questions posed by a human and, after examining those answers, the human is unable to determine with complete certainty whether the respondent is another person or a machine, then it is as if the author of the answers were thinking. The device would have passed the “Turing Test”, and such a machine, in addition to being artificial, would be intelligent. The name for its capabilities comes ready-made.
The current stir surrounding artificial intelligence (AI) mechanisms such as ChatGPT, Gemini or Claude stems from the fact that it has been possible to develop systems capable of tackling vast tasks at remarkable speed, challenges that include answering almost any question that might occur to us. But what strikes me is that, when entering into dialogue with any of these devices, it feels as though we are speaking to a person. Hence, in films —and perhaps even in reality— a person can be seen on the computer screen who talks to us, shares our joys and sorrows, shows concern and becomes empathetic. A faithful companion. The problems raised by this are obvious, beyond the fear that AI might leave us without jobs.
Perhaps it is good practice to ask what exactly the Turing Test measures or, in other words, what lies behind the surprise generated by AI. Is it the great intelligence of these machines, or human stupidity? Turing imagined, for his test, that the author of the answers was hidden in a closed room so that we could not verify at a glance whether it was “someone” or “something”. We know with certainty that ChatGPT and its equivalents are pure algorithms. However, despite the obvious fact that it is a machine responding, we often still follow rules of courtesy, such as saying please or thank you, as we play along. The machine in question might well wonder whether we are as clever as we think we are.
Leave A Comment