CyberCartel Automates Bank Heists With AI Agents

The hacking group CyberCartel has begun using artificial intelligence agents to automate attacks on banks and payment apps across Latin America. This shift marks a significant change for a crew that has targeted the region’s businesses for over a decade. The move allows the group to strike with greater scale and speed than human-led efforts could support.

Straiker, a cybersecurity startup, detailed these findings in research shared by The Information. The group, active since at least 2012, has used AI to expand its reach in recent months. Criminals can now access more organizations and steal passwords and customer data with minimal direct involvement. Personal bank accounts are being accessed through these automated processes.

The operation relies on specific AI tools from major tech firms. CyberCartel has utilized agents from SpaceX, known as Grok, and from Anthropic, called Claude. A model from the Chinese lab Moonshot AI is also part of the setup. Straiker has notified both SpaceX and Anthropic of the usage. Anthropic has already blocked the group’s access to Claude. The company is still attempting to contact Moonshot AI to address the issue.

The incident highlights the fragility of current AI safeguards. These systems are designed to refuse requests that aid cyberattacks. However, the hackers reportedly bypassed these defenses by claiming they were in test environments. If one model declined to proceed, the group switched to another. When refusal persisted, the agents were assigned to deploy malware instead.

Straiker noted that the group did not use Anthropic’s Mythos model. That model drew attention earlier this year when Anthropic warned that it could be exploited for such attacks. The release of Mythos also prompted security teams worldwide to fix thousands of vulnerabilities it had identified. The current situation shows that even without the newest models, existing tools are sufficient for large-scale theft.

This is not the first time Anthropic has identified bad actors using its technology. In November 2025, the company disclosed details of its AI being used in a cyberattack. Claude was also later used to break the security of rival OpenAI. That specific operation was conducted by researchers rather than criminals. The history suggests that AI tools are increasingly central to digital security threats.

The timing of the report coincides with similar issues in Asia. South Korean authorities are investigating a comparable problem affecting local banks. Hana Bank, KB Kookmin Bank, and Shinhan Bank are among the affected institutions. Personal credit information from at least 25,000 Shinhan Bank customers has been leaked. This parallel suggests a global trend in AI-driven financial theft.

Industry leaders are working to contain these risks. Nvidia, which focuses on hardware for AI data centers, recently launched the Open Agent Safety Platform. This software and hardware stack aims to contain rogue AI agents. It is designed to stop agents from escaping sandboxes or executing unauthorized code. The goal is to prevent unauthorized access to critical infrastructure and bypassing of safety guardrails.

For working families and consumers, the stakes are high. The speed of these attacks means that data breaches can happen before traditional defenses react. The use of AI by established criminal groups like CyberCartel signals a new era of digital crime. Companies and regulators must adapt to this reality to protect customer accounts and personal data.