Cybersecurity researchers at NorForge have uncovered extensive evidence of AI-assisted cyberattacks targeting businesses worldwide. The investigation, published on October 19, 2026, reveals how attackers are using automated AI agents to identify vulnerabilities, attempt unauthorized access, and compromise websites with limited human involvement.
According to Reuters, one operation targeted more than 1,200 websites, with over 100 apparent successful intrusions.
Exposed AI Dashboards Reveal Hacking Operations
Researchers discovered more than 20 publicly accessible control panels associated with Cairn, a platform that coordinates AI agents to perform complex tasks.
The exposed dashboards revealed targeted websites, attack instructions, discovered vulnerabilities, and operational results. Some also contained sensitive credentials, including passwords and API keys, highlighting security weaknesses in the attackers’ own infrastructure.
One particularly active operation, identified as AIA0001, involved 14 exposed panels and more than 1,100 attack projects. Its targets included online retailers and payment-related systems across the United States, Canada, Japan, and Australia.
How AI Agents Are Changing Cyberattacks
Cairn is a general-purpose AI platform rather than a tool specifically designed for hacking. It allows users to define objectives while autonomous agents carry out tasks, analyze results, and adjust their approach.
Although this technology can support legitimate security testing, attackers can exploit the same capabilities to automate vulnerability discovery and intrusion attempts.
NorForge estimated that the systems it observed achieved their objectives in approximately 10% of attempts, with successful projects taking an average of just under 23 hours.
Some tasks cost as little as $5 in AI processing expenses. Reuters estimated that the largest observed campaign consumed approximately $5,000 in AI resources, equivalent to roughly $50 per apparent successful intrusion.
Businesses Face a Growing Automated Threat
The investigation identified apparent compromises involving organizations such as Deere & Company, SharkNinja, and Cummins, although the extent of any data exposure remains uncertain
The findings illustrate how AI could lower the cost and technical barriers associated with cyberattacks, enabling smaller groups to target large numbers of websites.
For businesses, the development reinforces the importance of timely software updates, strong authentication, secure credentials, and continuous monitoring.
While AI also provides valuable tools for detecting and preventing cyber threats, its growing use in offensive operations is creating a more challenging security environment for organizations worldwide.
Leave a Reply