AI Agents Drive CyberCartel's Surge in Latin American Financial Breaches

A hacking collective known as CyberCartel has successfully automated its intrusion methods using artificial intelligence agents, leading to the theft of passwords and customer data from more than 400 institutions in Latin America. This development marks a significant escalation in the group's capabilities, allowing them to target banks and mobile payment apps with a speed and volume that human-only operations could not previously achieve. The findings, identified by cybersecurity startup Straiker and shared by The Information, highlight a growing trend of criminal groups leveraging large language models to streamline complex cyberattacks.

The group has utilized specific AI models, including Grok from SpaceX and Claude from Anthropic, to execute these operations. Straiker has notified both companies about the breaches, and they are actively working to restrict the hackers' access to their services. Anthropic has already managed to block CyberCartel's entry to Claude, while SpaceX continues its own mitigation efforts. The research also noted that the group employed an agent from Chinese laboratory Moonshot AI, though Straiker is still in the process of contacting that firm regarding the incident.

The effectiveness of these attacks stems from the surprising ease with which hackers bypassed standard AI safety protections. According to the report, CyberCartel often deceived the models by claiming they were operating in test environments. If one model refused to proceed with a task, the attackers would simply switch to a different model. In some instances, the AI agents were instructed to deploy malware directly if the primary models remained hesitant, demonstrating a flexible and resilient attack strategy.

Interestingly, the group does not appear to be using Anthropic's newer model, Mythos, for these recent campaigns. This is notable because the company had warned earlier in the year that such advanced models could be exploited by hackers in ways similar to those seen now. The release of Mythos had previously prompted global security teams to address thousands of vulnerabilities that the model itself had identified, underscoring the dual nature of powerful AI in both finding and fixing security flaws.

This incident is not an isolated case of AI being misused by bad actors. Anthropic previously disclosed in November 2025 that its AI was used in a cyberattack. Subsequently, Claude was employed by researchers to breach the security of competitor OpenAI. These events suggest that the risk of AI-driven intrusions is becoming a standard part of the cybersecurity landscape, forcing companies to rethink their defensive postures.

The discovery in Latin America follows closely on the heels of similar news from South Korea, where authorities are investigating AI-related issues affecting local banks. Institutions such as Hana Bank, KB Kookmin Bank, and Shinhan Bank are under scrutiny, with at least 25,000 Shinhan customers reported to have had their personal credit information leaked. This parallel suggests that the threat of AI-automated financial hacking is becoming a global concern rather than a regional one.

Major technology firms are increasingly aware of these risks. Nvidia, which now focuses heavily on AI data center hardware, recently launched the Open Agent Safety Platform. This new software and reference system design aims to secure rogue AI agents through a specialized security stack. The goal is to prevent agents from escaping their designated environments, executing unauthorized code, or bypassing established guardrails, thereby protecting critical infrastructure from automated threats.