AI agents are now running automated cybercrime factories
Original headline: Attackers are using AI agents to automate cybercrime at 'industrial scale,' FBI and AWS warn — theregister.com
More in Cybersecurity news
The FBI and Amazon Threat Intelligence have exposed a new breed of automated cyberattacks using open-source AI framework ARTEX. Attackers are deploying AI agents that autonomously steal API credentials for AI models at "industrial scale," targeting platforms like One API, Dify, and Ollama. These stolen keys then fuel more agents in a self-sustaining cycle, compromising cloud infrastructure to access managed AI services and secret stores.
The campaigns also include sophisticated deepfake pipelines designed to bypass identity verification at cryptocurrency exchanges. While one real-time deepfake system was rejected, another achieved 11.5 frames per second on consumer hardware—close to live video expectations. This marks a dangerous evolution where AI isn't just a tool but an automated workforce for cybercrime, forcing defenders to combat scalable AI-on-AI attacks.
What to do: Monitor AI model gateway and proxy platforms (One API, New API, Veloera, Sub2API, LiteLLM) and AI application builders (Dify, FastGPT, Langflow, RAGFlow, Open WebUI, LobeChat, Ollama) for unauthorized access using stolen.
Read the original at theregister.com
Satirical commentary on real tech news, summarised with AI. Read the original source for the facts. How ZeroDaily works
