A cell of “malicious actors” in Yemen used Claude, Anthropic’s artificial intelligence (AI), in an attempt to develop remotely piloted weapons, the American company announced on Thursday, September 11. The American company disclosed the information as Houthi rebels are mounting a large-scale offensive in Yemen, notably against the Yemeni government and its main ally, Saudi Arabia.
“We have identified a cell of malicious actors based in northern Yemen that was running three weapons development programs,” Anthropic said in a report explaining how the startup halted illegal operations originating from several countries.
These programs involved “a guided rocket using a smartphone-type flight computer with autopilot,” a ballistic missile with a target range of over 2,000 km, and a missile with a “hypersonic glide vehicle,” according to the report.
The cell used Claude to substitute for a team of engineers specialized in the creation of guidance, navigation and control software for flying weapons. “We have no evidence that they managed to deploy an operational device,” Anthropic noted, while stating that these actors had conducted a test firing of a guided rocket that failed.
The startup says it has banned the associated accounts as soon as it identified these illegal activities and shared its information with public and private partners to “reduce the risks posed by these individuals.”
Although Anthropic did not name the actors involved, it is likely they are pro-Iranian Houthi rebels who control a large portion of northern Yemen, including the capital Sanaa.
