The U.S. government has reportedly been using Anthropic's AI model 'Mitos' to audit government software, despite ongoing conflicts over the use of AI.
On July 6, Reuters cited multiple sources stating that the Cybersecurity and Infrastructure Security Agency (CISA), part of the Department of Homeland Security, is utilizing Mitos to examine government code repositories.
According to sources, CISA is scanning government code repositories with Mitos to identify security vulnerabilities that could be exploited by foreign intelligence agencies or cybercriminals. This task is being handled by CISA's 'Attack Surface Evaluation Team,' which conducts digital security assessments and penetration testing across government agencies.
Reuters noted that it could not confirm the extent of the government code examined or the nature and severity of the vulnerabilities found. However, two sources indicated that multiple vulnerabilities have already been identified during the audit.
Since early this year, Anthropic has been at odds with U.S. authorities. Based in San Francisco, Anthropic refused to remove safeguards preventing its AI from being used for autonomous weapons or domestic surveillance in February. Subsequently, the Department of Defense designated Anthropic as a company posing a risk to the supply chain.
This designation has typically been applied to foreign companies suspected of supporting espionage activities.
However, a court halted this action in March. Following this, Anthropic reportedly eased tensions by privately releasing Mitos, an AI model known for its exceptional performance in identifying and exploiting cybersecurity vulnerabilities.
Reuters assessed that this case demonstrates the U.S. government's proactive stance on adopting Anthropic's AI tools, despite conflicts with the White House. Earlier, Axios reported that the National Security Agency (NSA) had begun using Mitos as early as April.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.