Cybersecurity expert and Cognition CSO Alex Stamos warns that artificial intelligence (AI) has elevated the scale and speed of cyberattacks to an unprecedented level. According to Stamos, various actors, from state-sponsored groups to independent criminals, are using AI-boosted tools to simultaneously detect and exploit vulnerabilities.
Stamos states that open-source AI models (such as the Chinese-based ARTEX) have completely automated attack preparations (the kill chain) that used to take months. Complex operations that previously only elite hackers could perform can now be executed by ordinary users (script kiddies) using standard hardware (such as Macs). Chain attacks targeting AI service providers like Fireworks and Cursor serve as concrete examples of this new era.
Emphasizing that the cost of defense remains extremely high while the cost of attacks approaches zero, Stamos notes that the tech world must adopt new approaches to survive this period. Arguing that merely finding and patching bugs is no longer sufficient, the expert asserts that rewriting foundational software infrastructures like the Linux kernel (for instance, in memory-safe languages like Rust) would be a cheaper and safer solution. Stamos also adds that AI-powered autonomous defense agents are necessary, but they must operate under human supervision and with deterministic limits.