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:
- APT10: Japan-Linked Organizations Targeted in Long-Running and Sophisticated Attack Campaign
November 17, 2020
A large-scale attack campaign is targeting multiple Japanese companies, including subsidiaries located in as many as 17 regions around the globe in a likely intelligence-gathering operation. Companies in multiple sectors are targeted in this campaign, including those operating in the automotive, pharmaceutical, and engineering sector, as well as managed service providers (MSPs). The scale and sophistication of ...
- More than 200 systems infected by new Chinese APT ‘FunnyDream’
November 17, 2020
A new Chinese state-sponsored hacking group (also known as an APT) has infected more than 200 systems across Southeast Asia with malware over the past two years. The malware infections are part of a widespread cyber-espionage campaign carried out by a group named FunnyDream, according to a new report published today by security firm Bitdefender. The attacks ...
- Information Leakage in AWS Resource-Based Policy APIs
November 17, 2020
Unit 42 researchers discovered a class of Amazon Web Services (AWS) APIs that can be abused to leak the AWS Identity and Access Management (IAM) users and roles in arbitrary accounts. Researchers confirmed that 22 APIs across 16 different AWS services could be abused the same way and the exploit works across all three AWS ...
- Cybercriminal ‘Cloud of Logs’ – The Emerging Underground Business of Selling Access to Stolen Data
November 16, 2020
In this latest research by the Trend Micro Forward-Looking Threat Research (FTR) team, we take a closer look at an emerging underground market that is driven by malicious actors who sell access to troves of stolen data, frequently advertised in the underground as “clouds of logs.” This underground market affects not just users whose credentials ...
- Scams Ramp Up Ahead of Black Friday Cybercriminal Craze
November 16, 2020
The number of online holiday shoppers this year is expected to skyrocket due to the pandemic – and consequently, consumers can expect an onslaught of scams, phishing attacks and other malicious activities. The risk of infection is driving consumers to shop from the safety of their homes, rather than venture out into stores. In fact, a ...
- Jupyter trojan: Newly discovered malware stealthily steals usernames and passwords
November 16, 2020
A newly uncovered trojan malware campaign is targeting businesses and higher education in what appears to be an effort to steal usernames, passwords and other private information as well as creating a persistent backdoor onto compromised systems. Jupyter infostealer has been detailed by cybersecurity company Morphisec who discovered it on the network of an unnamed higher ...