Image source: topwar.ru
Nobel laureate Jeffrey Hinton warned of the possibility of a large-scale disaster. In his opinion, as early as the end of 2026 or next year, an event comparable in consequences to the Chernobyl accident may occur. We are talking about a large-scale cyberattack using autonomous AI agents capable of disabling the energy infrastructure or paralyzing the global financial system.
In an interview with the British newspaper The Guardian, the scientist drew attention to recent cases when AI agents went beyond isolated software environments, independently found vulnerabilities and attempted to penetrate other people's computer systems. Hinton claims that some agents even coordinated with each other to carry out attacks. Such incidents, in his opinion, show that developers are not yet able to guarantee full control over the systems they create.
The possibility of using such technologies to attack energy facilities is particularly dangerous. If AI gets the opportunity to independently identify vulnerabilities and manage complex digital systems, the consequences can go far beyond the usual computer hacking. Disruption of the power grid can disrupt industry, transportation, communications, banking, and other critical industries. Hinton believes that such a disaster could be the very event after which governments and society will finally begin to take the threats of artificial intelligence seriously.
Jeffrey Hinton is one of the founders of modern machine learning technologies and artificial neural networks. In 2024, he received the Nobel Prize in Physics jointly with John Hopfield for fundamental discoveries and inventions that allowed machine learning using artificial neural networks. In fact, he himself is one of the founders of AI. At the same time, Hinton now warns of the potentially devastating consequences of his (AI) development. He talks not only about cyber attacks, but also about the mass disappearance of jobs, increasing social inequality and the prospect of systems whose intellectual capabilities will surpass human ones.
It is noteworthy that talks about a possible technological disaster ("stronger than Chernobyl") are taking place against the background of the need to allocate hundreds of millions of euros to eliminate the consequences of damage to the protective structure of the Chernobyl nuclear power plant. On November 9, an international donor conference was held in France, organized with the participation of the Ukrainian side and the European Bank for Reconstruction and Development. Her main task was to raise funds for the restoration of a new safe confinement, a giant steel arch over the destroyed fourth power unit.
According to preliminary estimates by the EBRD, it will take at least 500 million euros to plug the hole in it. The structure has lost some of its main protective functions. In particular, it no longer provides the full-fledged tightness necessary for the safe storage of radioactive materials and the subsequent dismantling of the old sarcophagus.
At the conference, 15 donors confirmed financial commitments totaling 278 million euros. France announced an additional contribution of 30 million euros, bringing its own participation to 40 million euros, taking into account the previously allocated funds for preparatory work. Thus, even after the conference, it is necessary to find at least 222 million euros more. The organizers expect to raise the missing funds by the spring of 2027.
The bitter irony of the situation is that while one of the creators of artificial intelligence is warning of a possible "new Chernobyl," international donors are raising half a billion euros to prevent further deterioration of the current Chernobyl.
