AI and Nuclear Weapons: Are We Controlling the Technology That Poses a Threat.

AI and Nuclear Weapons: Should We Worry?

According to Vox: People have always feared artificial intelligence and nuclear weapons. Movies like The Terminator, where Skynet becomes sentient and launches nuclear missiles, and War Games, which nearly led to nuclear war due to misunderstanding, illustrate these fears. The new film by Kathryn Bigelow Detroit raises questions about the potential involvement of AI in a nuclear strike on Chicago.

Artificial intelligence is already embedded in our nuclear infrastructure, emphasizes Josh Keating from Vox on the program Today, Explained. “Computers have been part of this system from the very beginning,” he says. “Some of the first digital computers were used during the making of the atomic bomb in the Manhattan Project.” However, the details of how this works remain unclear.

Are There Really Grounds for Concern?

Should we worry? Maybe, but not because AI could rise against humanity, Keating is confident.

“There is a part in Detroit where they try to figure out what happened and whether AI was involved in the incident. Do these films have grounds for concern?”

Interestingly, throughout history, nuclear war has never occurred in combat. Technological advancements have consistently raised questions about nuclear weapons. For example, in the 60s, Strategic Air Command even issued a response to Dr. Strangelove and The Fourth Protocol. In the 80s, the television movie The Day After became a powerful impetus for the nuclear freeze movement.

Artificial Intelligence in Nuclear Infrastructure

Many advocates of integrating AI into nuclear command systems assure: “It won’t be Skynet”. General Anthony Cotton, the current commander of Strategic Command, supports more AI implementation. He noted that “there will be no WOPR in Strategic Command.” However, concerns arise about the current state of AI.

“Do the people working on nuclear weapons really understand AI?”

We still don’t know how exactly AI affects nuclear infrastructure. Many will be surprised to learn how outdated the previous systems were. Until 2019, communication relied on 1980s flash drives. As part of the ongoing modernization of nuclear artillery, upgrades to these systems are planned.

Disadvantages of Integrating AI into Nuclear Systems

Even the most advanced AI models can make mistakes. Another issue is the potential for external intervention. There is also a phenomenon that many researchers call automation bias, where people overly trust information from computers.

Experience has shown that technological failures can lead to potential nuclear disasters, and it is the human factor that has kept escalation at bay. For instance, in 1979, Zbigniew Brzezinski received a message about the launch of thousands of missiles from the USSR, but it turned out to be a system error.

“Whatever the technologies, humanity always saves us from the brink when they fail.”

Fear certainly plays an important role in preventing nuclear conflicts. The most frightening thing is that AI could increase the likelihood of nuclear weapons being used by humans.

Therefore, judging by the conclusions from Detroit, it is clear that AI is already part of our nuclear infrastructure and will likely remain so. Those who deny the possibility of creating reliable AI should not be dealing with nuclear weapons.