The RAND Corporation, a well-known U.S. military think tank, has warned that artificial intelligence (AI) could start a nuclear war that will destroy humanity as early as 2040, according to foreign media reports.
During the Cold War, a state known as "mutually assured destruction" maintained a precarious balance between the two superpowers. But Rand said advances in artificial intelligence in the coming decades could undermine the binding force of "mutually assured destruction" on nuclear-armed states.
The RAND Corporation consulted a number of experts before publishing the report. In the future, AI could encourage humans to make catastrophic decisions, the report said. For example, improvements in sensing technology could lead to the destruction of some retaliatory forces, such as submarines and mobile missiles.
In addition, AI may also lure one country into launching a "pre-emptive" strike against another in order to gain bargaining power, even if the country had no intention of launching an attack.
Andrew Lohn, co-author of the report, said: "Some experts worry that the growing reliance on AI could lead to a new set of catastrophic mistakes. There are extreme risks in using AI until the technology matures. It is a big risk and is likely to be betrayed. Therefore, maintaining strategic stability in the coming decades may be difficult. All nuclear powers should take this issue seriously to limit this nuclear risk.”
Rather than focusing on other threats such as autonomous aircraft, the report primarily highlights the risks that the use of artificial intelligence technology could pose to military decision-making.
The report also mentions a false nuclear alert in the former Soviet Union in 1983. At that time, the computer of a nuclear missile warning center in the former Soviet Union issued an attack alarm, which was later confirmed to be a false alarm.
Edward Geist, another co-author of the report, said: "The link between nuclear war and artificial intelligence is not a new topic. Historically, there has been an intersection between the two. In fact, many artificial intelligence The early research on it was all to support the military, or to have a military purpose.”
Before that, many industry leaders and experts have warned about the application of artificial intelligence in the military field. For example, more than 100 industry insiders , including Tesla ( 280.597 , -0.09 , -0.03% ) CEO Elon Musk, have appealed to the United Nations to block the development of lethal automated weapons.
Article source: Xinhua Viewpoint