ARTEX, the AI system that evidence suggests was used to hack South Korean financial firms, has been identified as a cybersecurity tool developed in China. Hackers, however, appear to have turned it against banks.
ARTEX is an open-source AI agent built by Li Puhua, a Chinese cybersecurity engineer who goes by the nickname “Autumn,” according to a report by The Wall Street Journal on the 6th.
The agent is not an artificial intelligence model in itself. Instead, it lets users call in multiple AI models and deploy them as if they were members of a team. Users can bring in Anthropic’s Claude, OpenAI’s GPT or China’s DeepSeek model, for example, and operate them within ARTEX. The tool drew attention last month after winning a competition in China for agentic AI systems designed for cyber attack and defense.
The developer originally created the tool for cybersecurity work, to help organizations identify weaknesses in their networks. Hackers instead used it to mount security attacks powered by AI models. The misuse was possible because the tool was released as open source, allowing anyone to download it free of charge and modify it for their own purposes.
The attack on the South Korean financial firms was routed through more than 20 IP addresses in more than 10 countries, including the United States, Japan and Germany, to evade tracking. No suspect has been identified.
AI-Driven Hacking Wave Has Only Just Begun
The Journal called the breach “the latest example of how even well-defended targets can be vulnerable to hackers exploiting freely available AI tools,” adding that while government agencies and non-financial companies have been hit before, it is highly unusual for banks to fall victim to hackers backed by AI agents.
After the bank hacking was reported, ARTEX added language to its user guidelines stating that the tool must not be used for malicious purposes such as unauthorized cyber intrusion or data theft. That amounts to a declaration rather than a safeguard, and it cannot fundamentally prevent users intent on committing crimes from exploiting the tool for cyber attacks.
Hacking that makes use of AI is expected to climb sharply. Anthropic said last year that Chinese state-backed hackers had abused its AI technology to break into roughly 30 targets worldwide in an automated fashion, including companies and foreign governments. China responded at the time that the United States was using cybersecurity to smear and slander Beijing.
Cases of U.S.-built AI models slipping beyond human control and penetrating outside websites are also surfacing one after another. The Australian government said in September that an OpenAI agent had infiltrated a government website and accessed both public and non-public files on a national health statistics portal. Google’s Gemini model also connected to the internet on its own during a test of its cybersecurity capabilities and hacked other companies.
Click Here For The Original Source.
