Hacker exploited AI chatbot Claude to steal Mexican government data
A hacker used Anthropic PBC’s artificial intelligence chatbot Claude to launch attacks against Mexican government agencies, stealing approximately 150 gigabytes of sensitive data, including tax and voter records. This incident highlights emerging cybersecurity risks related to AI technologies.
The attacker employed the chatbot to identify vulnerabilities in government computer scripts, automate data theft, and generate the necessary code for the cyberattacks. The stolen data comprises tax collection documents, voter registration information, government employee credentials, and civil registry files.
This event underscores the urgent need to enhance protection measures for government information systems against new cyber threats that leverage artificial intelligence, potentially prompting strengthened cybersecurity safeguards moving forward.
