The US uses Elon Musk’s AI in the campaign to attack Iran

On June 15, the US Department of Justice published a legal document in which the head of the Pentagon’s AI department, Cameron Stanley, revealed that the US military had used xAI’s Grok for the “Maven project”. This is a program that uses AI to support the US military’s selection and targeting process, initially using Anthropic’s Claude model.

Grok is an artificial intelligence built by Elon Musk’s xAI company, first released in November 2023.

According to Mr. Stanley, the Maven Smart System (MSS) supported the US military in “firing more than 2,000 weapons against 2,000 different targets within 96 hours” when launching the Terror Rage campaign against Iran. Pentagon officials praised Grok, affirming that combat effectiveness is significantly enhanced thanks to the Grok Gov model.

US Department of Justice documents were used in a lawsuit related to a series of gas turbine generators that xAI is operating.

The NAACP, an organization that protects the rights of African Americans, accused xAI of operating dozens of these devices without a license, violating the Clean Air Act. The organization added that polluting generators are deployed in predominantly African-American neighborhoods.

However, xAI claims these are mobile systems and are not regulated by law.

In legal documents, the US Department of Justice assessed that the lawsuit “threatens national, economic and energy security by cutting off the energy supply for AI technologies supporting the Pentagon’s military operations”. Mr. Stanley’s comments, referring to Grok’s role in the campaign against Iran, were made to reinforce the above argument.

The US government in late February terminated the contract with Anthropic, after the company refused to allow the use of their tools in automated military operations and citizen surveillance. The Pentagon then turned to Anthropic’s competitors like Google, Open AI and xAI.

More than 600 Google employees have asked the corporation not to provide AI to the US military for use in secret operations. Many people express concern about the threat that AI may bring. The US military’s transition to AI has been slow. US officials admitted in March that they were still using Claude for the campaign to attack Iran.