AI requires the collaboration of a “ human ” by the need to apply common sense
One of the big mistakes in adopting systems based on artificial intelligence and machine learning is believing that they are sufficient on their own, according to Ramon Trias, president of AIS Groupy, pioneer of the IA in Spain.
For him, the integration of technology and business must be done while keeping in mind that managers are at the head and those responsible for the integration of digitization and the techniques associated with the future of the company. . “You have to get to the point where the need leads to the medium, not the medium that sets the standard, which is the approach we see in many companies. For each challenge, for each problem, it is necessary to analyze which technique is most appropriate to obtain an optimal response ”.
According to Trias, the capabilities of AI are mythologized, “forgetting that these tools do not have meta-knowledge, that is to say common sense. They are very good at solving very limited problems, but today they are not able to handle generalization, ie they do not know how to apply rules from one environment to another ” .
The president of AIS Group underlines that human beings can generalize and improvise, “AI not yet” and reminds that all systems learn from history and unforeseen events have no place. “Hence the importance of integrating them into the expert’s knowledge.” For him, it is up to the intuition and reflection expert and the Machine Learning expert to manage an extremely high number of variables and dimensions. The two must complement each other. So your error rate can be divided by 10.
AI and the digitalization of the economy
Trias points out that digitization began in the late 1970s in sectors such as banking and industry and, to a lesser extent, retail. Today, digitization, robotization and automation are very present there, while in others they are only just beginning. “Digitization,” he says, “has brought about changes that are here to stay, but it is not yet as profound a transformation as expected”.
The president of AIS Group recalls that we are currently experiencing a boom in artificial intelligence applied to digitization processes, but that AI, machine learning and mathematical optimization are relatively old techniques, which are now emerging thanks to improvements and cheaper. hardware, which even enables SMEs to realize processes and solutions that were previously reserved for research centers. “It also indicates that digitization produces a cost adjustment by shifting activity from businesses to users. One example is digital banking, where certain tasks that an employee previously performed, such as transfers, are now performed directly by users ”.
For Trias, artificial intelligence takes on real value when it is integrated into the rest of organizations’ information systems. “You have to know how to integrate technology and business,” he admits. “You have to understand the applicability, the monetary benefit, but also the service. You need to have a very clear vision of the possibilities for change that digitalization, AI, robotics or IoT bring to the business. Some will apply it to reduce costs, others to increase sales, others to improve processes… This is why it is important to learn how to manage projects that involve these technologies and how they fit into value chains ”.