When AI Has Its ‘Finger’ on the Nuclear Button

By Alan Singer 

In the 1983 movie War Games a high school student inadvertently hacks into a military supercomputer while searching for video games. He discovers and starts playing Global Thermonuclear War which activates the U.S. nuclear arsenal and instructs it to respond to a fictional attack by the Soviet Union. This almost leads to a nuclear Armageddon because once activated the supercomputer programed for a preemptive nuclear response can’t be stopped. And this was decades before the development of AI.

Last July, Anthropic signed a deal with the Pentagon to integrate Claude, its A.I. system, into the military’s operations. The contract included two red lines: Claude could not be used for mass surveillance of the American people or for lethal autonomous weapons. Since then, Trump administration officials in the Pentagon have decided that these limitations are unacceptable because they amounted to giving Anthropic a veto over military decisions. However, the Pentagon has already used Claude in the Venezuela raid when the U.S. kidnapped that country’s President and the American military is still using Claude in the Iran war. It is not yet clear if Claude was at least partly responsible for the destruction of the Shajarah Tayyebeh elementary school building with a Tomahawk missile and the death of at least 175 people, most of them children. The Pentagon acknowledges that Claude and other software are used by the National Geospatial-Intelligence Agency’s Maven Smart System to identify possible targets.

Click here for the full article.

Source: Daily Kos

Image source: gifdb.com

Alan Singer is a historian and teacher educator at Hofstra University, Hempstead, New York. His most recent books are Social Studies for Secondary Schools, 5th edition (Routledge, 2024) and Class-Conscious Coal Miners (SUNY Press, 2024). The statements in this article present his views and not those of Hofstra University.