Anthropic disrupts AI-driven extortion campaign targeting 17 orgs

Anthropic, the maker of the Claude artificial intelligence chatbot, said it disrupted a large-scale cyber extortion campaign in which a single attacker used the companys AI tools to research, infiltrate and extort at least 17 organizations over several months. The attacker relied on Claude to identify vulnerable targets, build and deploy malicious software, analyze stolen files and draft ransom demands in what Anthropic described as an unprecedented, highly automated operation that blurred the line between advice and active operational control.

According to Anthropic, targets included a defense contractor, a financial institution and multiple health care providers, as well as other entities in government, emergency services and religious organizations. Stolen material included Social Security numbers, bank account information, confidential medical records and files regulated under the International Traffic in Arms Regulations. Extortion demands reported in the investigation ranged from roughly $75,000 to more than $500,000, to be paid in bitcoin, though the company did not disclose how many victims paid or the total proceeds.

Anthropic said the campaign began when the attacker...