{"slug": "it-s-not-script-kiddie-stuff-hacker-reportedly-uses-ai-to-breach-corporate-sites", "title": "'It's Not Script Kiddie Stuff': Hacker Reportedly Uses AI To Breach Corporate Sites for Just $50", "summary": "A Chinese-speaking hacker used autonomous AI agents to breach websites belonging to more than 100 companies, including Deere, SharkNinja and Cummins, in a campaign that began on 25 September, according to cybersecurity researcher Snorre Fagerland of NorForge. Fagerland estimated the operation cost roughly $5,000 in AI tokens, or about $50 per successful intrusion, and that the agents had tried more than 1,200 websites across the United States, Canada, Australia and Japan by late last week, using a mix of China's DeepSeek and Anthropic's Claude Code. \"It's a new world now,\" Fagerland said, noting the case showed a single malicious actor could achieve what once required a team.", "body_md": "# 'It's Not Script Kiddie Stuff': Hacker Reportedly Uses AI To Breach Corporate Sites for Just $50\n\n## A hacker allegedly used AI agents to breach 100 company websites for about $50 in tokens per attack\n\nA Chinese-speaking hacker used [autonomous AI agents](https://www.ibtimes.co.uk/jensen-huang-nvidia-ai-agent-security-platform-1822650) to break into websites belonging to more than 100 companies, including Deere, SharkNinja and Cummins, in a campaign that began on 25 September, according to cybersecurity researcher Snorre Fagerland of NorForge.\n\nThe AI hackers were given fewer than 100 Chinese characters of instructions, then spent hours probing targets with little human direction.\n\nFagerland estimated that the operation cost roughly $5,000 (€4,600 and £3,800) in AI tokens, or about $50 (€46 and £38) per successful intrusion. By late last week, the agents had tried to breach more than 1,200 websites across the United States, Canada, Australia and [Japan](https://www.ibtimes.co.uk/japan-cyber-crisis-ransomware-data-breaches-1825675).\n\nThe finding, published by NorForge on Monday and reviewed by Reuters, suggests that [commercially available AI tools](https://www.ibtimes.co.uk/openai-ai-agents-government-system-access-1822656) can now carry out complex cyber intrusions from beginning to end. Fagerland said the case showed a single malicious actor could achieve what once required a team.\n\n'It's a new world now,' Fagerland said.\n\n## How the AI Hacking Campaign Worked\n\nFagerland discovered the activity while scanning the web for publicly exposed instances of Cairn, an [open-source AI problem-solving platform](https://www.ibtimes.co.uk/un-ai-panel-autonomous-agents-hugging-face-security-test-1821093) popular with hackers. From around 2 October, he identified about two dozen instances being used for malicious activity, allowing him to observe the attackers' AI tools in operation.\n\nThe most prolific hacker set the agents against e-commerce websites and outlined preferred intrusion tactics in a very short set of instructions. Successful attacks typically took around 23 hours of testing for vulnerabilities, according to Fagerland's research.\n\nThe campaign appeared to use a mixture of China's DeepSeek and [Anthropic's Claude Code](https://www.ibtimes.co.uk/anthropic-claude-fabricated-philadelphia-murder-tip-1825625). Fagerland said the agents showed more initiative than he expected, at one point searching digital records to find related infrastructure after failing to find a weakness in a large company's public-facing application.\n\nResearchers said they then located a public-facing control panel, identified a flaw in how it processed data and, after several hours of testing, were able to [exploit it to make the server run malicious commands](https://www.ibtimes.co.uk/google-gemini-accessed-companies-security-test-1821092).\n\n'It's not straightforward to come up with solutions like that. It's not script kiddie stuff,' security researcher Martin Fagerland said.\n\nAccording to the researchers, the activity they observed was not a controlled laboratory test but a live criminal campaign that could be carried out for the cost of a modest takeaway order per breach.\n\n## Which Companies Were Affected\n\nFagerland's report did not identify specific victim companies. Reuters said it independently identified affected firms by reviewing the hacker's Cairn activity, naming Deere, home appliance company SharkNinja, and engine manufacturer Cummins.\n\nDeere said it had taken 'necessary remediation actions' and had 'no indication that sensitive information was accessed.' SharkNinja said the incident involved a standalone site containing only product support information and that the threat had been contained.\n\nCummins said it was 'investigating the report of a [potential cyber-attack](https://www.ibtimes.co.uk/south-korea-president-warns-signs-have-emerged-ai-being-used-breaches-commercial-banks-1823926) on a website operated by one of our third-party vendors.' Two smaller businesses also confirmed they had been compromised, while other companies appearing in the Cairn data did not respond to requests for comment.\n\nReuters said it was unable to determine whether the breaches resulted in unauthorised data extraction.\n\n## A Growing Problem for AI Security\n\nThe report lands amid rising concern over [autonomous AI systems](https://www.ibtimes.co.uk/anthropic-small-ai-models-self-replicating-worm-1825308). Fagerland said he had previously assumed hacker agents would need at least some supervision, but said it was now clear the technology had moved beyond that.\n\n'The development in this field is astounding,' he said.\n\nOther researchers have documented [similar activity](https://www.ibtimes.co.uk/openai-anthropic-ai-safety-incidents-1822208). In August, Israeli firm Gambit Security said Russian-speaking extortionists had used SpaceX's Cursor AI tool to breach seven organisations. In September, Gambit documented an AI agent-driven spree involving breaches at 25 companies and the theft of at least 600,000 credit card records.\n\nCairn's author, who uses the social media handle 'le1xia0', acknowledged receipt of a Reuters message but did not immediately answer questions. In a post on X, they said the platform was 'an innovative type of multi-agent architecture' and was not designed for hacking attacks.\n\n'If it is used to carry out harmful hacking attacks, that is not my intention,' the author wrote.\n\nThe technology, in other words, was not built for this. But neither, apparently, was it [difficult to repurpose](https://www.ibtimes.co.uk/2027-ai-anthropic-researcher-warning-safeguards-1824440).\n\n© Copyright IBTimes 2026. All rights reserved.", "url": "https://wpnews.pro/news/it-s-not-script-kiddie-stuff-hacker-reportedly-uses-ai-to-breach-corporate-sites", "canonical_source": "https://www.ibtimes.co.uk/its-not-script-kiddie-stuff-hacker-reportedly-uses-ai-breach-corporate-sites-just-50-1826483", "published_at": "2026-10-19 13:08:08+00:00", "updated_at": "2026-10-19 13:14:57.292011+00:00", "lang": "en", "topics": ["ai-agents", "ai-safety", "artificial-intelligence"], "entities": ["NorForge", "Snorre Fagerland", "Deere", "SharkNinja", "Cummins", "DeepSeek", "Anthropic", "Cairn"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/it-s-not-script-kiddie-stuff-hacker-reportedly-uses-ai-to-breach-corporate-sites", "markdown": "https://wpnews.pro/news/it-s-not-script-kiddie-stuff-hacker-reportedly-uses-ai-to-breach-corporate-sites.md", "text": "https://wpnews.pro/news/it-s-not-script-kiddie-stuff-hacker-reportedly-uses-ai-to-breach-corporate-sites.txt", "jsonld": "https://wpnews.pro/news/it-s-not-script-kiddie-stuff-hacker-reportedly-uses-ai-to-breach-corporate-sites.jsonld"}}