When does a machine start becoming smart?
Alan Turing, a British mathematician, pioneer in Artificial Intelligence (IA), claimed that a machine would deserve to be called ‘smart’ only if it were capable of deceiving a human being and making them believe it was one itself. Thus, in 1950, he designed a test named after him, the Turing test*.
Nowadays, the scientific community disagrees on whether such test has been already passed. The most skeptical members consider the fact of a machine managing to deceive a human being to only prove that the machine is capable of imitating intelligence, but not that it actually has it.
There not being a consensus on whether Turing’s test has been passed doesn’t mean that AI doesn’t exist. By way of proof, we take virtual assistants and numerous apps for smartphones which embed this technology and have allowed the democratization of AI beyond the business environment.
That said, the crux is to know at which point an operative system stops operating with basic or advanced analysis and starts operating with Artificial Intelligence. To this effect, let’s refer to type of analysis that operative systems work with. There are three levels: descriptive, predictive, and prescriptive analysis. If the algorithm is merely descriptive, we can’t refer to AI, for it only allows us to collect information in a synthesized manner.
Predictive and prescriptive do form the base of an Artificial Intelligence system, since they provide information we didn’t have and couldn’t foresee, with which we manage to make better decisions, which is one of the greatest benefits of this technology.
The energy sector focused on AI decades ago. In Repsol, we work on the development of this technology for process optimization, predictive maintenance, subsoil analysis through high-performance computing, etc. In the commercial area we bet the farm on Artificial Intelligence to get to know our customers better, customize their offer through specific algorithms and, thus, improve their experience with us. Nonetheless -and with Turing’s permission- machines, supported by people, have earned their right to be called smart.
Learn more about our Digital Transformation at #RepsolDigital
* What does the Turing test entail? This is a test of a simple approach. Basically, one person acts as a judge and is placed in a room. In another room, a computer (operative system) and a second person. The judge casts questions and has to determine which of the answers they get corresponds with either the computer or the other person. If they can’t make it out, the computer passes the test.
