AI, the end of the world

I am transcribing a paragraph from an article I read: "90% of office jobs will be destroyed or transformed until they become unrecognizable in the next ten or fifteen years. This new technology threatens administrative staff, intermediaries, technicians, and managers. It will do to intellectual work what machines did to factories. Yes, the prediction is catastrophic because almost all of us already work, in one way or another, in these types of jobs."The text could belong to any of the articles published this week about AI. But no. Do you know what year it is from? It was written by Tom Peters in Time magazine on May 22, 2000. Twenty-six years ago. The technology he was referring to was the internet and cloud software.In recent weeks, several researchers have left their AI companies. Jacob Coxon resigned from Anthropic accusing the company and OpenAI of "playing with our lives." Other experts speak of loss of control, machines capable of disobeying us, and a possible extinction of humanity. There are even those who calculate the probability that we will not survive the next decade.These warnings are important. Especially when they come from people who know the technology from the inside. But listening to them does not oblige us to accept their conclusions.What is more dangerous: an AI or an atomic bomb? The bomb can destroy a city in a few seconds. We have been living with thousands of these weapons for eighty years, not because they are harmless, but because we have created treaties, protocols, authorization systems, and deterrence mechanisms.Something similar will happen with AI. Its security will not depend on the fact that it is incapable of disobeying, but rather that human beings build limits, firewalls, and controls to prevent it. Every technology first develops its possibilities and then the rules intended to contain them. It takes us time, but we end up doing it.Will AI destroy jobs? Of course. Just like agricultural mechanization, the textile industry, automotive, computers, and the internet. At the same time, however, it will create activities that we do not yet know how to name today and will skyrocket our capacity to research, diagnose, learn, and produce.There will be cyberattacks, errors, fraud, and perverse uses. The Internet also brought them. What great innovation has not caused accidents or been used to do harm? Risk demands control; it does not invalidate an entire technology.Apocalyptic predictions have an advantage: if they do not come true, no one asks for explanations. In the year 2000, the internet was supposed to end work. Today we work with the internet. In a few years, we will discover that AI did not come to replace humanity either, but rather to force it to take responsibility for a new form of power while benefiting from it.