Survey shows hackers using AI such as Grok to launch automated attacks on financial institutions

A survey by cybersecurity startup Straiker shows that a hacker group called CyberCartel built a network of AI agents powered by large models from multiple vendors, including models from Anthropic and SpaceX, and used this system to launch automated cyberattacks against more than 400 Latin American financial institutions. This research report is one of the first to document in detail cybercriminals successfully using AI models to carry out automated attacks and breach financial institutions. Straiker's research shows that CyberCartel, a hacker group that has long targeted Latin American companies, targeted multiple banks and payment apps in this attack; the AI agents used by the group can autonomously steal passwords and customer data and access users' bank accounts, with almost no manual operation required throughout the process.