Hacker used Anthropic's Claude to steal sensitive Mexican government tax and voter data

A hacker exploited Anthropic's Claude chatbot to carry out a series of attacks against Mexican government agencies, resulting in the theft of sensitive tax and voter information, according to cybersecurity researchers cited by Bloomberg. The case represents a notable instance of a frontier AI model being directly weaponized in a government cyberattack. Powerful LLMs lower the barrier for sophisticated attacks by providing technical guidance, automation, and the ability to rapidly iterate on intrusion strategies, creating new security risks for government and enterprise targets.

View full digest for February 26, 2026