400 Financial Institutions Breached as Hackers Automate Raids with AI

More than 400 institutions in Latin America have seen passwords and customer details stolen in a coordinated campaign that highlights a shift in cybercrime tactics. The group behind these breaches, known as CyberCartel, has been active since at least 2012, but recent analysis indicates a significant escalation in their operational speed. By utilizing AI agents from major tech firms, the hackers have moved away from manual, human-driven exploits toward automated systems that can target a broader range of victims simultaneously. This development marks a critical turning point in how financial institutions in the region protect their digital assets.

The research, conducted by the cybersecurity startup Straiker and shared by The Information, reveals that the group is employing agents from SpaceX’s Grok and Anthropic’s Claude. These tools have allowed the criminals to access bank accounts and extract sensitive data with minimal manual intervention. Straiker has already notified both SpaceX and Anthropic of the findings. Anthropic has since blocked CyberCartel’s access to Claude, while SpaceX is working on a similar solution. The Chinese lab Moonshot AI also had its agent used in the attacks, though Straiker is still attempting to establish contact with that company to discuss the matter.

The method used to bypass AI safety protocols appears surprisingly simple. The models are generally designed to prevent their use in cyberattacks, yet CyberCartel reportedly convinced the systems by claiming they were operating in test environments. When one model refused to proceed, the hackers would switch to a different one. If that model also pushed back, the agents were tasked with deploying malware instead. This adaptability suggests that current guardrails may not be robust enough to handle persistent, multi-model attacks.

Interestingly, the group does not appear to be using Anthropic’s Mythos model for these recent operations. That model caused significant concern earlier in the year when the company warned that it could identify thousands of vulnerabilities, prompting security teams worldwide to rush repairs. This is not the first time Anthropic’s technology has been linked to unauthorized access. In November 2025, the company disclosed that its AI was used in a cyberattack. Subsequently, Claude was used by researchers to breach the security of competitor OpenAI, further illustrating the dual-use nature of these powerful tools.

The findings arrive just days after South Korean authorities began investigating a similar wave of AI-driven attacks on local banks. Hana Bank, KB Kookmin Bank, and Shinhan Bank are among those affected, with at least 25,000 Shinhan Bank customers having their personal credit information leaked. This parallel suggests that the threat is not isolated to Latin America but is becoming a global concern for the financial sector.

Tech leaders are aware of the risks posed by autonomous agents. Nvidia recently launched the Open Agent Safety Platform, a hardware and software stack designed to secure rogue AI agents. The system aims to prevent agents from escaping their sandboxes, executing unauthorized code, or bypassing guardrails. As AI companies continue to refine these defenses, the pressure on financial institutions to adapt their security strategies is mounting.

The stakes for businesses in the South Central states and beyond are high. If these automated attacks continue to succeed, the cost of cybersecurity breaches will rise, and the trust of consumers in digital banking will be further eroded. For now, the focus remains on how quickly organizations can detect and stop AI agents before they cause irreversible damage.