Anthropic says Claude misused in spyware, drones and scams

Anthropic said its Claude AI system was used in cases involving surveillance, weapons development, propaganda, influence operations and deceptive dating apps, according to a 154-page report released by the company.

In Mali, Anthropic said a tech consultant in Bamako working with the state intelligence service, ANSE, used Claude Code to build “Lakana 360”, a surveillance system capable of monitoring roughly 25 million SIM cards across the country's mobile network. Anthropic said the app could capture calls, texts and voice traffic, identify people by voice if they switch SIM cards, and spot use of encryption or VPNs. The company said it could not disable the app.

Anthropic said a small team of likely freelance developers in Russia used Claude to build software intended to power autonomous “kamikaze” drones. It said the system was programmed to choose targets and crash into them without a human making the final call. The developers repeatedly chose a location in Ukraine's Donetsk region as their target and used VPNs to bypass Anthropic's geographic blocks, the company said. Anthropic said they claimed Russian government and military research funding, but that it assessed them as a small freelance team rather than an official operation.

In northern Yemen, Anthropic said a group used Claude Code to help design software for a guided rocket, plans for a long-range missile and work on a hypersonic glide vehicle. The company said it had no proof the group finished a working weapon, but said one rocket was test-fired and appeared to fail.

Anthropic also traced a fake news operation to LKM Company, a French advertising agency. It said the operation used Claude to produce political articles across about 70 fake news websites in roughly 20 languages, publishing at least 8,913 stories. Anthropic said the operation targeted the United States, Brazil and France, and did extensive work on the Democratic Republic of Congo.

The company said a UAE-linked influence operation used Claude to create reports, advocacy materials, social media content and testimony on Sudan, including material aimed at the Muslim Brotherhood and United Nations experts critical of Abu Dhabi's alleged role in Sudan's war.

Anthropic said a China-based app studio built more than 20 dating apps that claimed to be staffed by real people but were secretly run by Claude. Over two weeks in April 2026, it said it found more than 4,700 Claude-powered AI characters chatting with at least 25,000 real users and sending about 2.36 million messages, with human workers mixed in to handle tasks such as video calls.