A recent report by the artificial intelligence company Anthropic revealed that an Iran-linked group used American-developed AI technology to target U.S. naval forces during the U.S.-Iran conflict. The group, referred to as a "threat actor" in the report, utilized a large language model called Claude AI to gather and analyze open-source information—such as publicly available satellite imagery and social media posts—to identify the locations of U.S. naval assets and recommend targeting strategies. After the account was identified and banned, Anthropic shared the findings with the U.S. government, helping to disrupt the threat. Experts warn that such scenarios may become more frequent as adversaries increasingly use AI to improve their military and strategic operations. AI models like Claude are designed to quickly process and analyze vast amounts of publicly available data, which can be a powerful tool for both beneficial and harmful purposes. Nikita Shah, a senior fellow at the Center for Strategic and International Studies, explained that adversaries are using AI to enhance their existing tactics, making them more effective, efficient, and faster. Sam Bresnick, a research fellow at Georgetown University's Center for Security and Emerging Technology, noted that this has made it increasingly difficult for the U.S. military to conceal its activities from potential threats. Although access to Claude is blocked in Iran, adversaries can still create accounts using virtual private networks (VPNs) and foreign phone numbers. A July report by cybersecurity firm Recorded Future suggested that AI may have accelerated Iran's existing cyber and military capabilities. It also pointed to the possibility that AI-driven tactics, which were refined during the conflict in Ukraine, played a role in recent Iranian drone attacks against Israel and other countries in the Persian Gulf. In addition to military operations, Iran has used AI for "soft war" strategies, such as spreading propaganda through social media. Anthropic found that three Iranian state-aligned accounts used Claude AI to manage digital campaigns, translate content, and create posts that appeared to come from independent news sources. These posts were shared across platforms like X (formerly Twitter), TikTok, and Instagram, aiming to influence public opinion and criticize the U.S. and Israel. This kind of AI-assisted influence campaign is not unique to Iran, as other adversaries are also using AI to enhance their operations and spread disinformation. Experts note that while the U.S. has maintained an advantage in situational awareness, the rapid advancement of AI models is allowing other countries to close the gap. The Pentagon is working to counter this by developing AI-based countermeasures, such as creating digital "personas" that mimic real-life behavior to obscure the digital traces of military activities. While a Pentagon official declined to comment on intelligence matters, experts agree that the national security community has been preparing for the possibility that adversaries will use AI to refine, scale, and amplify their operations. Anthropic's report underscores the growing challenge that tech companies face as they navigate the complex and evolving landscape of adversarial threats.