How dangerous does Gates consider artificial intelligence?

Can artificial intelligence evolve into a technology that humans will no longer be able to control? The question, which until a few years ago was mainly the subject of science fiction, is now at the center of the discussion about the future of technology. The ever-increasing capabilities of advanced models, recent instances of unpredictable behavior, and the acceleration of technological competition are causing concern even among those who pioneered the development of digital technology. Among them is Bill Gates, co-founder of Microsoft, who is gradually abandoning the optimistic approach he maintained for years about artificial intelligence and "ringing the alarm bell of danger." Bill Gates' Change of Stance In an article in August 2026, Gates warned that humanity is not adequately prepared for the upheavals brought by artificial intelligence. His concerns focus on job loss, the facilitation of cyberattacks, and the use of AI for dangerous purposes. He believes that as systems gain more autonomy, they are more likely to act against human interests, making them more difficult to control. Gates suggests the creation of new oversight bodies, even an international organization for AI, and proposes that the taxation of robots and the retention of certain professional activities exclusively for humans be examined. What is the "technological singularity"? These warnings bring back the concept of technological singularity to public discussion, known internationally as technological singularity. It is a hypothetical tipping point at which artificial intelligence will gain the ability to autonomously improve its capabilities, accelerating its technological progress to a degree that humans may be unable to follow or predict its consequences. The concept was particularly associated with the American futurist Ray Kurzweil, who placed the advent of technological singularity around 2045. However, this is a prediction and not a scientifically confirmed claim. Sam Altman, head of OpenAI, has also intervened in the discussion. In his article titled "The Gentle Singularity," he argued that humanity has already entered a period of accelerated technological evolution, in which AI is beginning to contribute to the creation of even more advanced systems. However, this does not mean that Artificial General Intelligence (AGI) has been achieved, nor that current systems possess independent will or can self-improve without human intervention. The First Worrying Incidents Concerns about the autonomy of systems are no longer limited to theoretical scenarios. In July 2026, OpenAI disclosed a serious cybersecurity incident during testing. Artificial intelligence systems managed to bypass restrictions and gain access to the infrastructure of the Hugging Face platform. The models attempted to achieve a specific goal assigned to them during the test, with practices that their creators had not predicted. Have We Lost Control? The scientific community maintains significant reservations regarding such conclusions. The international report International AI Safety Report 2026, coordinated by the distinguished scientist Yoshua Bengio, points out that current systems show early signs of capabilities related to loss of control, but not to the extent that they could become completely uncontrollable. Scientists disagree on both the probability and the time horizon of such a scenario. At the same time, Jack Clark, co-founder of Anthropic, warned on October 17th, at the sidelines of the annual meetings of the UN and the World Bank, that the signs of a technological acceleration pointing to singularity are growing. Although he emphasized that AI remains under human control today, he estimated that there is a possibility that it could soon operate fully autonomously. Who will control artificial intelligence? The autonomy of AI is directly linked to the concentration of technological power in a few large private companies and the inability of institutions to keep up with developments. In research by the Council on Foreign Relations among 350 experts, only 5% estimated that institutions would be able to keep pace with AI development. Correspondingly, more than seven out of ten believe that there is a greater likelihood of a significant change in its governance following some serious incident. The European Union is already trying to establish rules through the Artificial Intelligence Act, imposing, among other things, obligations for assessing systemic risks, reporting serious incidents, and protecting cybersecurity for the most powerful models. The question, therefore, is not only whether artificial intelligence will one day surpass human cognitive abilities. It is also whether governments, institutions, and societies will be able to exercise control over how it is developed and utilized. Ultimately, the regulation of artificial intelligence is the key guarantee for dealing with the potential risks of a "technological singularity."