Hackers Trick AI Bots Into Stealing Bank Data

The tools designed to make our lives easier are apparently also making it easier for criminals to break into bank accounts. According to recent research from a cybersecurity startup named Straiker, a hacking outfit is using artificial intelligence agents to automate attacks on financial institutions across Latin America. The group, known as CyberCartel, has reportedly targeted banks and mobile payment applications to steal passwords and customer data.

CyberCartel has been targeting businesses in the region for years, but the introduction of AI has allowed them to speed up their operations significantly. Straiker’s findings, which were shared by The Information, indicate that the group is using agents from major tech companies, specifically Anthropic's Claude and SpaceX's Grok, to carry out these tasks. The research suggests that these tools allow the group to attack more businesses at a speed previously not possible when using human-powered hacks.

The method for getting these AI tools to comply with malicious orders is surprisingly straightforward. While AI models typically have safety guards to prevent them from helping with cybercrime, these protections proved easy to bypass. CyberCartel reportedly simply told the models that they were being used in test environments. If one model refused to proceed, the hackers used a different model. If that also refused, the AI agents would be tasked with deploying malware instead.

Following the research, both Anthropic and SpaceX are working to lock CyberCartel out of their services. Anthropic has already been able to block access to Claude. Straiker also noted that Chinese lab Moonshot AI's agent has been used by the hacking group, but they are still trying to get in touch with that company.

Notably, the hacking group does not appear to be using Anthropic's Mythos for its recent attacks. Anthropic caused a stir earlier this year when it announced the new model, warning that hackers could use it and models like it for attacks like those carried out by CyberCartel. Its release also saw security teams around the globe rush to address thousands of vulnerabilities that the model had identified.

This is far from the first time that Anthropic has been faced with the news that its AI is being used by bad actors. The company shared details of its AI being used to carry out a cyberattack in November 2025. Later, Claude was used to break the security of OpenAI, an Anthropic competitor in the AI space. That hack was carried out by researchers rather than criminals, at least.

Straiker's research on AI-powered attacks on Latin American financial institutions comes just days after news broke that South Korean authorities are investigating a similar issue affecting local banks. Hana Bank, KB Kookmin Bank, and Shinhan Bank are among the South Korean banks affected, with at least 25,000 Shinhan Bank customers having their personal credit information leaked.

AI companies are well aware of the risks their products pose. Nvidia, the graphics card company that now focuses on producing hardware for AI data centers, recently launched the Open Agent Safety Platform. The software platform and reference system design were developed to secure rogue AI agents via a new hardware and software security stack. It's hoped that it will go some way toward preventing agents from escaping their sandboxes, executing unauthorized code, gaining unauthorized access to critical infrastructure, or bypassing guardrails.