Illustration=ChatGPTA China-linked hacking group used artificial intelligence (AI) agents to carry out cyberattacks against more than 200 organizations worldwide, and Korea-based institutions involved in foreign policy were among the targets, according to new findings.Adam Meyers, head of counter adversary operations at U.S. cybersecurity firm CrowdStrike, said in an online media briefing on the 17th that “the China-linked hacking group ‘VAPOR PANDA’ used an AI-based tool called ‘ARTEX’ to target at least one organization in Korea.”Meyers, referring to think tanks and foreign policy–related institutions as Korea-linked targets, said, “We can confirm that some of the attacks succeeded.” He did not disclose specific organization names or the extent of damage. Asked whether the affected organizations were notified, he said, “We notified multiple incident response teams and informed the relevant organizations directly or through those teams.”Meyers said, “This operation is separate from the ARTEX attacks that targeted Korea’s financial sector early last month.” CrowdStrike assesses that while both campaigns used ARTEX, the actors behind them were different.CrowdStrike said in a report published on the 29th of last month that VAPOR PANDA used ARTEX to attack more than 200 organizations worldwide. The attacks, carried out from the 17th to the 29th of last month, targeted foreign ministries in Latin America, Europe, Asia, and the Middle East; pharmaceutical, semiconductor, defense, and telecommunications companies; and think tanks.VAPOR PANDA is a China-linked hacking group that CrowdStrike has tracked since 2021. It previously focused on government agencies and companies in the energy and resources sectors, but it has recently expanded its scope by leveraging AI tools.Adam Meyers, head of CrowdStrike’s incident response operations, speaks at an online media briefing held on the morning of the 17th. /Courtesy of ZoomARTEX, used in the latest operation, is an AI-based penetration testing tool built on large language models (LLMs) that repeatedly scans for vulnerabilities in web applications and application programming interfaces (APIs) and uses information gathered at each step to conduct follow-on attacks. VAPOR PANDA primarily ran ARTEX using the Chinese AI model DeepSeek, and it also obtained paid subscriptions to other cutting-edge AI models as secondary options.In one intrusion, roughly 1.4 billion input tokens were consumed. Meyers said, “AI is replacing tasks that humans used to perform, swapping labor hours for token costs,” adding, “If they run their own language models, they could also significantly reduce inference costs.”Automation via AI is also lowering the bar for hacking. ARTEX looks for known vulnerabilities such as SQL injection and cross-site request forgery and uses information obtained at each stage for the next attack. “Tasks that human operators used to do are now automated by AI, enabling less-skilled attackers to use them,” Meyers said. “Sophisticated groups like VAPOR PANDA are deploying this at scale across hundreds of targets.”There were also cases where previously stolen data was reused in AI-driven attacks. VAPOR PANDA fed directory information it obtained in 2022 from a European government agency into ARTEX this year to support attacks. By pairing accumulated tradecraft and data with AI, the group increased its operational efficiency.On AI-driven cyberattacks, Meyers said, “What matters is speed and scale. We have observed this type of intrusion for years.” He added, “Companies should apply appropriate security technologies across various operating systems, including Windows, Linux, and Mac, and ensure visibility into threat detection and their cloud environments. If you establish security controls and continuously hunt for threats, you can detect and block AI-enabled attacks.”※ This article has been translated by AI. Share your feedback here.
Click Here For The Original Source.
——————————————————–
