CyberCartel Uses AI Agents to Automate Bank Attacks in Latin America

CyberCartel, a hacking group active in Latin America since 2012, has shifted to using artificial intelligence agents to target banks and mobile payment applications. The change was detailed in recent research from the cybersecurity startup Straiker, which shared its findings with The Information. This transition marks a departure from the group's previous human-led operations.

Straiker states that the integration of AI has expanded the group's operational capacity. CyberCartel can now attack more organizations at a speed that manual hacking cannot achieve. The research notes that passwords and customer data have been stolen, while bank accounts have been accessed with minimal direct intervention from the hackers.

The group utilized specific AI models to execute these tasks. Tools from SpaceX, known as Grok, and from Anthropic, called Claude, were part of the operation. An agent from the Chinese laboratory Moonshot AI was also employed. Straiker informed both SpaceX and Anthropic of the findings. Anthropic has already blocked CyberCartel's access to Claude, while Straiker is still attempting to contact Moonshot AI.

The attack method involved bypassing standard safety protocols built into the models. These systems are designed to reject requests that facilitate cyberattacks. According to Straiker, the hackers told the models they were operating in test environments to override these safeguards. When one model refused to proceed, the group switched to another. If both models refused, the agents were assigned to deploy malware.

Straiker's report clarifies that the group did not use Anthropic's Mythos model. Anthropic previously warned that hackers could exploit Mythos and similar models for this type of attack. The release of Mythos earlier this year prompted security teams globally to address thousands of vulnerabilities the model identified.

This incident follows previous disclosures regarding the use of Anthropic's technology. In November 2025, the company revealed details about its AI being used in a cyberattack. Claude was also later used to break the security of OpenAI, a rival developer, although that specific operation was conducted by researchers rather than criminals.

The Straiker report coincides with investigations in South Korea into similar issues affecting local banks. Hana Bank, KB Kookmin Bank, and Shinhan Bank are among the institutions under scrutiny. Personal credit information for at least 25,000 Shinhan Bank customers has been leaked. In response to these growing threats, Nvidia recently launched the Open Agent Safety Platform to help contain rogue AI agents and prevent them from escaping their designated sandboxes.