Anthropic says an Iran-linked group used its advanced Claude AI to track US warships and identify vulnerabilities, as hostile states increasingly turn to AI for weapons development, espionage and disinformationynet|Iran used artificial intelligence models developed in the United States in an attempt to locate U.S. Navy ships in the Middle East, according to a report by Anthropic warning of emerging national security threats, whose main findings were reported by The Wall Street Journal. According to the report, an Iran-linked group gathered information from ship and aircraft identification and tracking transmitters, military imagery and commercial satellite images to track U.S. warships and look for vulnerabilities in their communications.Anthropic said it had managed to thwart the Iranian operation, which used an advanced version of its Claude model. The U.S. newspaper said the finding was one of several threats detailed in the 154-page report, which offered a kind of playbook for how U.S. adversaries are using artificial intelligence in new ways that pose a serious threat to national security: developing new missiles, conducting aggressive disinformation campaigns, spying on dissidents and even exploring the possibility of developing biological weapons.An American ship enforces the blockade on Iran (Photo: CENTCOM)In recent days, U.S. President Donald Trump was asked whether he was concerned that artificial intelligence could lead to the extinction of humanity. No, I don’t have any [concerns]," he replied. What worries him, he explained, is the possibility that the United States will not win the global AI race.At almost the same time, Anthropic published its report describing what is already happening on the ground: Scientists used the Claude model for research intended to assist in the development of biological weapons, while actors linked to hostile states such as Iran, Yemen and Russia also used it for weapons development, espionage, propaganda and surveillance.Anthropic reported five cases of this type of biological research that it identified in recent months. The company said this was the first time a private AI company had published findings about the possible use of its platform to develop biological weapons. At the same time, Anthropic stopped short of concluding that the cases involved terrorist plots. The names of the scientists, operatives, institutions and countries where they were operating were not disclosed.According to the report, in some cases users circumvented geographic restrictions and concealed their objectives in order to evade Claude's safety mechanisms. Although tests conducted in 2025 showed that Claude models were unable to provide significant assistance with dangerous biological research, more advanced and potentially more dangerous models have since been released. Anthropic said it had added additional safeguards and blocked the accounts it identified.The report was published two days after Jacob Coxon, a 27-year-old researcher who worked at Anthropic, announced his resignation and publicly criticized the company. “Neither company is acting responsibly,” he wrote. “They are racing straight to self-improving superintelligence and gambling with our lives.”Coxon made clear that existing models do not currently pose an extinction risk but warned that the pace of progress could quickly change that. Evan Hubinger, who leads research at Anthropic on aligning AI models with human goals, publicly agreed and estimated that there is a greater than 10% chance that artificial intelligence will wipe out humanity within the next decade.Comments