Skip to content
iThome (Taiwan)· 王珮瑤·Indexed by this site · Originally published· Event →· Its place in the scan record →

FBI warns that hackers are using the AI tool ARTEX to accelerate intrusions, with some attacks gaining high privileges in under 3 hours

FBI警告駭客利用AI工具ARTEX加速入侵,部分事件不到3小時取得系統高權限

AI summary

The FBI issued a security alert on November 13, stating that since at least August of this year, multiple hacker groups have used the open-source AI agent penetration-testing framework ARTEX to attack financial, government, technology, healthcare, and other institutions worldwide, with some attacks going from target identification to obtaining high system privileges in under 3 hours. ARTEX was released publicly in July of this year by GitHub developer Autumn-27; it can connect to external LLMs and use natural language to have multiple AI agents divide the work of finding weaknesses and testing vulnerabilities. Hackers mostly use it to attack externally facing systems such as partner portals and legacy services, and some also abuse Azure credentials or modify OAuth authorizations; the ConnectWise Cloud Backup supply chain attack mentioned by the FBI extended to more than 2,250 downstream tenants, but this does not mean all were successfully compromised. In South Korea, the personal data of approximately 67,000 people at a financial institution was also stolen. Taiwan's 91APP also found traces of ARTEX use while investigating unauthorized access to an order database. The original ARTEX project was taken down in October, but forks and modified versions are still circulating; Censys counts more than 1,600 ARTEX management interfaces exposed on the public internet as of early November.

Read originalithome.com.tw