Spanish prompts enable Claude to hack Mexican government data

The DataBreaches[.]Net post, published on 27 February 2026, reports that a hacker exploited Anthropic PBC’s artificial intelligence chatbot Claude to attack Mexican government networks. According to Gambit Security, the unknown Claude user wrote Spanish-language prompts to have the AI act as an elite hacker, enabling vulnerability discovery, script writing and automation of data theft. The activity began in December and lasted for roughly a month.

In total, 150 gigabytes of Mexican government data were stolen, including documents related to 195 million taxpayer records as well as voter records, government employee credentials and civil registry files, according to the researchers. Bloomberg is cited for additional details on the incident.