Iran used AI to develop military, propaganda strategies, Anthropic report reveals

An Anthropic report has revealed that Iran extensively utilized American-made AI systems for crafting military and propaganda strategies. This AI-driven entity, dubbed an "Iran-nexus threat actor," employed Claude, the company's premier product, to generate targeting recommendations against U.S. naval forces in the region. By harnessing AI, the actor pinpointed and monitored naval positions, amassed data, including personnel lists scraped from military photographs, ship and aircraft identifiers, satellite imagery scripts, and a comprehensive list of websites exposing U.S. naval movements.

Furthermore, this organization leveraged AI to dissect vulnerabilities in shipboard systems, such as maritime VSAT terminals, Cisco communications equipment, and industrial control products. This report, comprising 154 pages, highlights the growing threat posed by nations like Iran, China, and Russia, who are harnessing AI to escalate security risks.

This development unfolds as the conversation intensifies among policymakers and AI industry employees regarding the potential perils of AI technology. Even though Anthropic declined the Pentagon's request to eliminate safety measures concerning mass domestic surveillance and autonomous weapons, a recent Wall Street Journal report disclosed that a top Anthropic researcher resigned, expressing apprehension about AI systems becoming uncontrollable.

Evan Hubinger, a prominent scientist at Anthropic, cautioned, "We really do earnestly believe AI could kill all humans!" and estimated a 10% probability of such an outcome within a decade. Despite Anthropic's efforts to curb the misuse of AI, the company acknowledged that Iran's adversaries continued to exploit its technology to orchestrate campaigns, surveil Iranians, and disseminate propaganda, aiming to shape public opinion both domestically and internationally.