Hackers Used Claude and Grok to Automate Attacks on More Than 400 Latin American Targets

A hacking group known as CyberCartel reportedly used AI agents powered by models from Anthropic and SpaceX to automate attacks on banks and payments companies across Latin America. For PSPs and financial institutions, the point is not the headline-grabbing brand names; it is that automated attack workflows are now being used against customer data and bank accounts at scale.

1. According to a report Wednesday, Oct. 21, from The Information, citing research by cybersecurity startup Straiker, the hackers used a web of AI agents powered by models such as Anthropic’s Claude and SpaceX’s Grok.
2. The research says CyberCartel attacked more than 400 targets in Latin America, including banks and payments companies. The tools were used to steal password and consumer data and to access customer bank accounts autonomously, with little human effort.
3. The Information described the findings as one of the first detailed reports showing hackers using AI models to automate breaches into financial institutions. The operational detail matters: attacks can now be executed quickly and in greater numbers, which can overwhelm corporate defenses.
4. Straiker CEO Ankur Shah told The Information: “Offense is way ahead of defense right now, that’s just how it is. Enterprises are just not ready.” That is the practical takeaway for banks and PSPs deciding where to spend on fraud and security controls.
5. Anthropic said it began an investigation after Straiker disclosed the findings and quickly blocked the hackers’ use of Claude. The report also notes that many companies are now investing heavily in AI-powered cyberprotections.

For financial institutions, the relevant context is already inside PYMNTS Intelligence’s own research: in September, close to 80% of financial institutions said the complexity of new technology systems was their chief barrier to innovation, compared with 48% one year earlier. The same report said behavioral analytics, device fingerprinting and enhanced knowledge-based authentication reached 100% adoption among surveyed institutions this year, while machine learning and AI were used by 96%.