AI’s control over nuclear weapons may trigger a catastrophic war

New York: The periodical, Bulletin of Atomic Scientists, has warned that handing over control of nuclear weapons to Artificial Intelligence (AI), is an invitation to an all destructive catastrophic war, as seen in the movie, Terminator. In the last few years, leading countries like the United States, China and Russia were observed to gear up their nuclear preparedness once again. At the same time, the use of AI is on the rise in the defence sector while researchers and experts from around the world have consistently warned against it terming it dangerous.

Three experts from Cornell University in the United States have published a research paper regarding AI and nuclear weapons. The article highlights the growing use of AI in the defence sector, which has now extended to nuclear weapons as well. The report mentions, the system, which would provide instructions to launch an attack to the unit handling nuclear weapons, was automated and AI-based.

Countries such as Russia and China are making extensive use of AI to surpass the United States, but experts have warned that it posed a dangerous threat. The US experts issued a severe warning that said, ‘Leading countries in the defence sector may be drawn towards the advanced AI technology as they believe it is a safe alternative. However, it poses a risk of causing harm, which cannot be understood unless a major disaster occurs.’

The report also mentions the ‘Poseidon’ nuclear torpedo Russia has developed. The US experts have expressed concerns over the AI-based torpedo which could trigger a series of such defence systems. An incident in Russia from 1983 was mentioned to emphasise why nuclear weapons should not be under control of the AI.

The report emphasises on the need for human control over nuclear systems. It states, ‘In the then Soviet Union, Lieutenant Colonel Stanislav Petrov, was assigned the critical responsibility concerning nuclear weapons. He had received warnings on the US missiles from the early warning system, operational at the time which even included some audio-visuals. However, Lieutenant Colonel Petrov ignored the messages, using his decision-making prowess. The concerned messages were proved to be a false warning, and a nuclear war between the United States and the USSR was averted.’

Three months ago, US Deputy Secretary for Defence Robert Work had issued a severe warning which asserted that AI’s control over nuclear weapons would prove very dangerous. He raised concerns by asking, ‘What will happen if China has a nuclear command based on AI and it ordered a pre-emptive strike after issuing messages of the United States initiating an attack?’

In the last few years, many countries have deployed critical AI-based systems in their defence forces. There are significant debates over the issue on the international forum regarding the regulations and other related matters. The United Nations (UN) is handling the subject, and an independent group was set up for the purpose.