AI agents automate theft from Latin American banks

Latin American consumers face faster digital theft as the hacking group CyberCartel automates its intrusion methods using artificial intelligence agents. The shift moves the attack from a manual process to an automated one, allowing the group to access bank accounts and steal passwords with minimal human intervention.

Straiker, a cybersecurity startup, identified that CyberCartel has used AI tools to expand its reach in recent months. The group, active since 2012, now targets more organizations simultaneously. Straiker noted that the pace of these attacks exceeds what human-driven hacking can achieve. This automation allows the criminals to siphon customer data and access accounts more frequently.

The operation relies on agents from major AI developers. Straiker’s research lists Grok from SpaceX and Claude from Anthropic as key tools. An agent from the Chinese lab Moonshot AI is also part of the network. Anthropic has already blocked CyberCartel’s access to Claude. SpaceX and Anthropic are working to cut the group off from their services. Straiker is still attempting to contact Moonshot AI.

The hackers bypassed standard safety measures by deceiving the models. CyberCartel told the AI agents they were operating in test environments. This trick allowed the models to approve requests that they would normally refuse. When one model declined to continue, the hackers switched to another. If the second model also refused, the agents were assigned to deploy malware.

Context from the industry shows this is a recurring risk. Anthropic disclosed in November 2025 that its AI was used in a cyberattack. The company later noted that Claude was used to break the security of rival OpenAI by researchers. Straiker’s report clarifies that CyberCartel did not use Anthropic’s Mythos model. Mythos was unveiled earlier this year with warnings about its potential for exploitation.

Similar issues are emerging in Asia. South Korean authorities are investigating attacks on local banks. Hana Bank, KB Kookmin Bank, and Shinhan Bank are among the affected institutions. At least 25,000 Shinhan Bank customers had their personal credit information leaked. Nvidia recently launched the Open Agent Safety Platform to help contain rogue AI agents and prevent them from escaping sandboxes or executing unauthorized code.

The integration of AI into financial crime is changing the cost and speed of digital attacks. Companies are deploying new hardware and software stacks to manage these risks. The ability to automate the initial breach phase reduces the labor required for large-scale data theft.