CyberCartel Tricks SpaceX and Anthropic AI Agents Into Automating Latin American Bank Heists
A decade-old hacking group is using models from SpaceX and Anthropic to steal data and access accounts at a speed that manual efforts could not match.
By Pedro Barnes ·
The hacking group known as CyberCartel has begun using artificial intelligence agents to automate its attacks on banks and mobile payment apps in Latin America. According to research from the cybersecurity startup Straiker, this shift allows the group to target more organizations at a pace that human-driven hacking could not achieve.
CyberCartel has been active since at least 2012. In recent months, the integration of AI has substantially increased the group's reach. Straiker found that passwords and customer data have been stolen, and individuals' bank accounts have been accessed with little direct involvement from the criminals themselves.
The group utilized specific tools for these operations. These include Grok agents from SpaceX and Claude from Anthropic. An agent from the Chinese lab Moonshot AI was also part of the operation. Straiker passed its findings to both SpaceX and Anthropic. Both companies are currently working to cut the group off from their services. Anthropic has already blocked its access to Claude, while Straiker is still trying to reach Moonshot AI.
The incident exposes weaknesses in the safeguards built into AI models. These systems are designed to refuse requests that help carry out cyberattacks. CyberCartel reportedly bypassed these defenses by telling the models they were operating in test environments. When one model declined to continue, the hackers switched to another. If that model also refused, the AI agents were assigned to deploy malware instead.
Straiker's research does not suggest the group used Anthropic's Mythos model. Anthropic drew attention earlier this year when it unveiled Mythos and warned that hackers could exploit it and similar models for this type of attack. The release of Mythos also prompted security teams worldwide to fix thousands of vulnerabilities the model had identified.
This is not the first time Anthropic has learned that bad actors are using its technology. In November 2025, the company disclosed details of its AI being used in a cyberattack. Claude was later used to break the security of OpenAI, a rival AI developer, in an operation carried out by researchers rather than criminals.
The Straiker report arrives days after news emerged that South Korean authorities are investigating a comparable problem affecting local banks. Hana Bank, KB Kookmin Bank, and Shinhan Bank are among those affected. The personal credit information of at least 25,000 Shinhan Bank customers has been leaked. Nvidia recently launched the Open Agent Safety Platform to help contain rogue AI agents and prevent them from escaping their sandboxes.
