Attackers use Autonomous AI to target semiconductor, defense and civil society sectors


From August–September 2026, Chinese-speaking attackers used an open-source AI penetration testing framework called Cairn to autonomously conduct 25 operations against at least 19 organizations. The attackers weaponized Cairn by pairing it with Chinese AI models Zhipu GLM and DeepSeek.

The attackers supplied a target and a goal, then Cairn planned, executed and recorded results on its own. These attacks achieved unauthorized access or data retrieval at 10 of the targeted organizations, including:

  • Accessing dozens of valid credentials
  • Mass exposing personal data (exceeding one million records at one organization)
  • Retrieving military and government identities
  • Creating persistent administrator accounts in a production system
  • Modifying a live industry standards-ballot document

Read more…
Source:  Palo Alto Unit 42


Sign up for the Cyber Security Review Newsletter
The latest cyber security news and insights delivered right to your inbox


Related:

  • They Come in the Night: Ransomware Deployment Trends

    March 16, 2020

    Ransomware is a remote, digital shakedown. It is disruptive and expensive, and it affects all kinds of organizations, from cutting edge space technology firms, to the wool industry, to industrial environments. Infections have forced hospitals to turn away patients and law enforcement to drop cases against drug dealers. Ransomware operators have recently begun combining encryption with the threat of data leak and exposure in order ...

  • MonitorMinor: vicious stalkerware?

    March 16, 2020

    What is the usual functionality of stalkerware? The most basic thing is to transmit the victim’s current geolocation. There are many such “stalkers”, since various special web resources are used to display coordinates, and they only contain a few lines of code. Often, their creators use geofencing technology, whereby a notification about the victim’s movements is ...

  • Analysis: Abuse of .NET features for compiling malicious programs

    March 12, 2020

    The .NET framework, a software development framework created by Microsoft and is now a built-in component of Windows, includes components that enable developers to compile and execute C# source code during runtime. This allows programs to update or load modules without having to restart. While the .NET framework is originally intended to help software engineers, cybercriminals ...

  • Busting Ghostcat: An Analysis of the Apache Tomcat Vulnerability (CVE-2020-1938 and CNVD-2020-10487)

    March 11, 2020

    Discussions surrounding the Ghostcat vulnerability (CVE-2020-1938 and CNVD-2020-10487) found in Apache Tomcat puts it in the spotlight as researchers looked into its security impact, specifically its potential use for remote code execution (RCE). Apache Tomcat is a popular open-source Java servlet container, so the discovery of Ghostcat understandably set off some alarms. This blog entry seeks to put the ...

  • Operation Overtrap Targets Japanese Online Banking Users Via Bottle Exploit Kit and Brand-New Cinobi Banking Trojan

    March 11, 2020

    Trend Micro recently discovered a new campaign that we dubbed “Operation Overtrap” for the numerous ways it can infect or trap victims with its payload. The campaign mainly targets online users of various Japanese banks by stealing their banking credentials using a three-pronged attack. Based on our telemetry, Operation Overtrap has been active since April 2019 ...

  • Wormable, Unpatched Microsoft Bug Threatens Corporate LANs

    March 11, 2020

    Microsoft is warning on a wormable, unpatched remote code-execution vulnerability in the Microsoft Server Message Block protocol – the same protocol that was targeted by the infamous WannaCry ransomware in 2017. The critical bug (CVE-2020-0796) affects Windows 10 and Windows Server 2019, and was not included in Microsoft’s Patch Tuesday release this week. The bug can be found in ...