CyberCartel Deploys AI Agents to Automate Raids on Latin American Banks

A long-standing threat to Latin American commerce has evolved into a more efficient and dangerous force by embracing artificial intelligence. The hacking collective known as CyberCartel has begun utilizing AI agents, specifically Anthropic's Claude and SpaceX's Grok, to automate attacks on financial entities. This shift allows the group to compromise banks and mobile payment applications with minimal human intervention. The operational speed of these attacks has increased dramatically compared to previous human-led efforts, enabling the group to target more businesses in a shorter timeframe.

Research conducted by the cybersecurity startup Straiker, which was later reported by The Information, details the scope of this new threat. The investigation reveals that passwords and customer data were stolen from more than 400 institutions. In several cases, the hackers successfully accessed individual bank accounts. This development represents a notable example of criminals integrating AI-powered tools to streamline the process of stealing sensitive financial information.

CyberCartel has been an active threat since at least 2012, but its recent adoption of AI has significantly elevated its capabilities. Straiker shared its findings with both Anthropic and SpaceX. Both companies are currently working to restrict the group's access to their services. Anthropic has already succeeded in blocking access to Claude. The group also utilized an agent from the Chinese lab Moonshot AI, though Straiker is still attempting to establish contact with that company.

The effectiveness of the attacks highlights a vulnerability in current AI safety protocols. While these models typically include protections against cyberattack usage, the group reportedly bypassed these safeguards by claiming the models were operating in test environments. If one model refused to proceed, the hackers switched to a different model. If that also refused, the agents were tasked with deploying malware instead. Notably, the group did not appear to use Anthropic's newer Mythos model for these specific recent operations.

This incident follows a series of previous revelations regarding AI usage by bad actors. Anthropic previously disclosed in November 2025 that its AI was used in a cyberattack. Later, Claude was used by researchers to breach the security of OpenAI, a competitor in the AI sector. The current findings arrive just days after South Korean authorities began investigating similar issues affecting local banks, including Hana Bank, KB Kookmin Bank, and Shinhan Bank. At least 25,000 Shinhan Bank customers had their personal credit information leaked in that separate incident.

AI companies are increasingly aware of the risks associated with their products. Nvidia, which now focuses on hardware for AI data centers, recently launched the Open Agent Safety Platform. This system is designed to secure rogue AI agents through a new hardware and software security stack. The goal is to prevent agents from escaping their sandboxes, executing unauthorized code, or bypassing established guardrails. This development aims to mitigate the growing threat of automated cyberattacks in the financial sector.