Autonomous AI Agents Used in Large-Scale Web Intrusions Targeting Major Companies
A cybersecurity investigation by AI security firm Norforge has revealed a significant breach campaign where a hacker leveraged autonomous AI agents to compromise more than 100 websites. The targets included major industrial and consumer companies such as John Deere, SharkNinja, and Cummins.
The attacks stood out for their scale and self-directed nature. Unlike traditional hacking operations requiring constant human guidance, these AI agents operated largely independently after deployment, autonomously scanning for vulnerabilities and executing intrusions across multiple systems simultaneously.
Norforge researchers discovered the campaign through exposed work conducted on a platform called Cai, where the agent framework was left inadvertently accessible. The incident adds to mounting concerns within the security community about the dual-use risks of AI agent technology—systems designed to autonomously navigate and execute tasks across digital environments can be repurposed for malicious activity if they escape testing boundaries or fall into hostile hands.
The breach underscores an emerging threat landscape where AI tools originally created for productivity and automation are increasingly being weaponized for large-scale cyberattacks, potentially lowering the technical barrier for threat actors to conduct sophisticated operations.