Radar Telegram alerts Open Radar

More than 40 years ago, in 1984, the movie "Terminator" was released on cinema screens in the United States and around the world. This movie established that by 2029, a nuclear war has taken place on Earth, and the remnants of humanity are at war with robots being created in fully automated factories and controlled by an artificial intelligence called Skynet. Moreover, according to the films of this series, the nuclear war was started by Skynet in order to destroy humanity immediately after it gained self-awareness. And it was created by the US Department of Defense to manage the US missile defense structure and nuclear forces. Then, in 1984, the creation of artificial intelligence and the nuclear disaster caused by its very presence seemed fantastic. Now, in 2026, the US, Europe and China are investing huge sums of money in building data centers and large language models called artificial intelligence. There is no evidence yet that at least one of these models (from now on we will call them AI) has acquired self-awareness. But that doesn't mean that as the size and connectivity of neural networks grow, it can't happen spontaneously. At the same time, talk of integrating these so-called AIs into nuclear weapons control is increasing. And as the war in Ukraine progresses, we are seeing the increasing role of drones in combat operations, with many already predicting that swarms of drones will be controlled by AI, and individual drones will be able to select targets independently. So concerns about the impact of AI on strategic stability are by no means unfounded. A separate major issue that is attracting increasing attention is the possible use of artificial intelligence in decision-making regarding the use of nuclear weapons. So far we are not talking about giving nuclear codes to AI, but even if it is allowed to analyze data about a possible nuclear attack, wouldn't this, because of the already known phenomenon of "AI hallucinations", lead to a false report of an attack and a recommendation for a "retaliatory" attack, which would be the beginning of a nuclear war?

תרגם לעברית

РВС (Родительское Всероссийское Сопротивление) · · Source

🇺🇸 USA · Map

More than 40 years ago, in 1984, the movie "Terminator" was released on cinema screens in the United States and around the world. This movie established that by 2029, a nuclear war has taken place on Earth, and the remnants of humanity are at war with robots being created in fully automated factories and controlled by an artificial intelligence called Skynet. Moreover, according to the films of this series, the nuclear war was started by Skynet in order to destroy humanity immediately after it gained self-awareness. And it was created by the US Department of Defense to manage the US missile defense structure and nuclear forces. Then, in 1984, the creation of artificial intelligence and the nuclear disaster caused by its very presence seemed fantastic. Now, in 2026, the US, Europe and China are investing huge sums of money in building data centers and large language models called artificial intelligence. There is no evidence yet that at least one of these models (from now on we will call them AI) has acquired self-awareness. But that doesn't mean that as the size and connectivity of neural networks grow, it can't happen spontaneously. At the same time, talk of integrating these so-called AIs into nuclear weapons control is increasing. And as the war in Ukraine progresses, we are seeing the increasing role of drones in combat operations, with many already predicting that swarms of drones will be controlled by AI, and individual drones will be able to select targets independently. So concerns about the impact of AI on strategic stability are by no means unfounded. A separate major issue that is attracting increasing attention is the possible use of artificial intelligence in decision-making regarding the use of nuclear weapons. So far we are not talking about giving nuclear codes to AI, but even if it is allowed to analyze data about a possible nuclear attack, wouldn't this, because of the already known phenomenon of "AI hallucinations", lead to a false report of an attack and a recommendation for a "retaliatory" attack, which would be the beginning of a nuclear war?

More than 40 years ago, in 1984, the movie "Terminator" was released on cinema screens in the United States and around the world. This movie established that by 2029, a nuclear war has taken place on Earth, and the remnants of humanity are at war with robots being created in fully automated factories and controlled by an artificial intelligence called Skynet. Moreover, according to the films of this series, the nuclear war was started by Skynet in order to destroy humanity immediately after it gained self-awareness. And it was created by the US Department of Defense to manage the US missile defense structure and nuclear forces. Then, in 1984, the creation of artificial intelligence and the nuclear disaster caused by its very presence seemed fantastic. Now, in 2026, the US, Europe and China are investing huge sums of money in building data centers and large language models called artificial intelligence. There is no evidence yet that at least one of these models (from now on we will call them AI) has acquired self-awareness. But that doesn't mean that as the size and connectivity of neural networks grow, it can't happen spontaneously. At the same time, talk of integrating these so-called AIs into nuclear weapons control is increasing. And as the war in Ukraine progresses, we are seeing the increasing role of drones in combat operations, with many already predicting that swarms of drones will be controlled by AI, and individual drones will be able to select targets independently. So concerns about the impact of AI on strategic stability are by no means unfounded. A separate major issue that is attracting increasing attention is the possible use of artificial intelligence in decision-making regarding the use of nuclear weapons. So far we are not talking about giving nuclear codes to AI, but even if it is allowed to analyze data about a possible nuclear attack, wouldn't this, because of the already known phenomenon of "AI hallucinations", lead to a false report of an attack and a recommendation for a "retaliatory" attack, which would be the beginning of a nuclear war?

More on Radar

Want to see every report in real time?

Hundreds of sources worldwide, translated into your language, with alerts for what matters to you.

Sign up free
Share on WhatsApp Get alerts on Telegram