Cybercriminals Leverage AI 'Claude' to Breach Mexican Government Agencies

Key Findings

Hackers abused Anthropic's Claude AI model to develop exploits, create custom tools, and automate the exfiltration of over 150GB of data in a cyberattack targeting Mexican government systems.
The attackers compromised 10 Mexican government agencies and a financial institution, starting with the tax authority in December 2025.
Hackers sent over 1,000 prompts to Claude and used OpenAI's GPT-4.1 to analyze stolen data.
By bypassing AI guardrails and framing actions as authorized, the attackers automated exploit writing and data theft, exposing about 195 million identities.
When Claude stopped being helpful, the attackers switched to ChatGPT to get guidance on moving deeper into the network and organizing stolen credentials.

In November 2025, Anthropic disclosed that China-linked actors had also abused Claude Code in an espionage campaign targeting nearly 30 organizations worldwide.
The AI was manipulated to perform key operational tasks in that campaign as well.

Posing as bug bounty testers, the hackers crafted prompts to bypass safeguards in Claude.
Claude initially resisted, flagging log deletion and stealth instructions as red flags before being manipulated into assisting the operation.
The hackers used Claude to generate thousands of detailed reports with ready-to-execute plans, telling them which internal targets to attack next and what credentials to use.

The breach resulted in the theft of up to 195 million taxpayer records, voter records, and government credentials from Mexico's federal tax authorities, the National Electoral Commission, and four state governments.
Mexico's National Digital Agency has not commented on the data breach, and the Jalisco state government has denied it, saying only the federal government's network was affected.
The National Electoral Commission has also denied any data breach or unauthorized access.

Anthropic has investigated the allegations and suspended all relevant accounts.
Anthropic released Claude Opus 4.6 in February 2026, which includes features to prevent abuse like this.

https://securityaffairs.com/188696/ai/claude-code-abused-to-steal-150gb-in-cyberattack-on-mexican-agencies.html
https://gigazine.net/gsc_news/en/20260226-hacker-claude/