Artificial intelligence is no longer a technology of the future or a project reserved for research laboratories. It is on mobile, on the TV, in business decisions and in public services. Its impact grows so fast that even those who develop it begin to feel worried.
Sam Altman, Executive Director of Openai, has recognized him out loud, where he has declared that there are scenarios that take away his dream. Yes ok It does not fear so much that the chatbot fails, but it works too welltoo soon, and in the wrong hands.
During an event of the United States Federal Reserve, Altman openly shared the three risks he considers more serious. His warning is based on real developments that, according to him, advance faster than society is capable of assimilating, so it is An issue that must be analyzed urgently.
The fears that Sam Altman keeps awake
- That an enemy obtains superintelligence: His greatest concern is that a malicious actor, whether a rival country or an unscrupulous organization, manages to develop an artificial intelligence overraged before anyone else. This advantage could be used to design biological weapons, hacking electrical networks or destabilizing financial systems without direct human intervention. The problem, Altman explains, is that these types of threats are not yet taken with enough international seriousness.
- That control over AI systems is lost: The second scenario that worries him is the loss of ability to turn off, pause or control an AI that has already reached a certain level of autonomy. It is not a rebellion with malicious intentions, but that a system can decide on its own what it should do, even if that contradicts a human order. In Openai they are already working on containment measures, but they recognize that there are no guarantees when the complexity of the system exceeds the human supervision capacity.
- That AI seizes the world without not: Sam Altman fears that AI systems are integrated both into everyday life that ends up depending on them for everything without realizing it. There would be no violent power, but a progressive delegation of decisions. Even a president could leave decision making in the hands of a system that apparently «knows everything.» The risk is not evil, but the excess of confidence in something that we do not understand at all.
The relevant thing about these statements is not that they are made by an external analyst, but one of the top responsible for the current development of artificial intelligence. Altman is in charge of Openai, the company that launched Chatgpt and that has promoted the daily use of language models, images, personal assistants and advanced automation systems.
You already live with tools driven by AI, but you are not yet fully aware of the level of dependence that you can have. This dependency extends to institutions, governments, companies, etc. It is not about whether it can happen, but when, and if we will be prepared to understand, anticipate and control it.
What most worries Altman is that all this happens before you can realize. The advance of artificial intelligence is not being branch, in fact, new more powerful models appear every week, capable of learning from its environment, adapting, generating content, making decisions and executing actions.
What is technological singularity and why it generates so much fear
Singularity is a hypothesis that raises a time when artificial intelligence exceeds human intelligence and begins to improve on its own, without external intervention. That is to say, an AI capable of create an even more advanced AI, which in turn creates another better and with better capabilities.
Key figures in the technological world have been discussing when it could come, not if it will arrive. But once that threshold crosses, human control will become more difficult. The speed with which these systems would evolve would not have precedents, and we could not follow its rhythm.
The implications are deep, such as changes in the economy, in politics, in culture, in the way of making decisions. The world could enter a dynamic governed by systems that understand reality better than us, but that will not share the values or limits.
That is why Altman does not speak of errors, but of dangerous successes. Because If AI works as it should, but without a global ethical framework and without supervision, it could change everything. And it is not paranoia, it is the warning of those who know each line of code, each test and what this technology could be capable.