The Claude artificial intelligence, developed by the American company Anthropic, is being used or attempted to be used by various states and groups for military, propaganda and cyber purposes. According to a new report by Anthropic, Russia has tried to use Claude to develop navigation technology for drone swarms used in the war against Ukraine. In other cases, China has tried to use it for electronic warfare technology, while in Yemen it has been used for missile guidance systems. The report cites a total of six cases where Claude has been used to develop or improve weapons, including drones, missiles and guided munitions.
The report shows that dangerous uses of artificial intelligence are not just related to far-fetched scenarios where machines could go out of control. Today, AI is being used for concrete activities such as cyberattacks, propaganda, fraud and surveillance.
Anthropic says it has seen a sharp increase in attempts to misuse Claude in 2026. The company breaks these cases down into several categories, including the use of AI for weapons, disinformation, political interference, surveillance, cyberattacks, financial fraud, and potentially dangerous biological research. One of the main concerns is the use of AI for propaganda and election interference. According to the report, actors affiliated with Russia, Iran, and other countries have used Claude to create and distribute political content in multiple languages. In some cases, campaigns have been created that aim to influence public opinion or manipulate electoral processes.
The report also cites China-linked efforts to use artificial intelligence to surveil dissidents and minorities, including Uyghurs. In other cases, the technology has been used to monitor communications or gather intelligence. In the cybersecurity field, Claude has been used in attempts to breach computer network defenses, steal data, and conduct espionage. The report cites Russia-linked attacks on Ukrainian institutions, as well as attempts to target governments, diplomatic, and defense interests in Europe and the U.S. Another concern relates to the use of AI in biology. Anthropic says it has found no clear evidence that bioterrorists are using Claude to develop biological weapons, but the company is investigating several cases that have raised concerns.
Meanwhile, Anthropic founder Dario Amodei has called for the development of more advanced artificial intelligence models to be temporarily slowed down to allow more time for security mechanisms to be created. According to him, the risk comes not only from an AI that could become more intelligent than humans, but also from the way people are using this technology today.