North Korean Threat Groups Operationalize AI for Phishing, Malware, Agentic Cyber Operations: Microsoft
Microsoft Threat Intelligence says Jade Sleet, Coral Sleet and Jasper Sleet have operationalized AI, while Onyx Sleet and Pearl Sleet use the ARTEX framework.

North Korean threat actors are expanding their use of artificial intelligence to increase the speed, efficiency and complexity of cyberattacks, according to Microsoft Threat Intelligence.
Microsoft assessed that the groups Jade Sleet, Coral Sleet and Jasper Sleet have operationalized AI across their operations, the company said in a report.
The company also assessed Diamond Sleet, Onyx Sleet, Pearl Sleet and Sapphire Sleet as being at the accelerating stage of AI adoption, and Emerald Sleet as being at the expanding stage.
Onyx Sleet created more than 100 PDF lure files in August and September 2026 with target-specific details, and metadata indicated dozens of files were created within minutes, Microsoft said. The automation supported Onyx Sleet’s high-volume phishing campaigns against South Korean users.
Jade Sleet conducted job-themed targeting from June 8 to June 12, 2026 against individuals across cybersecurity, AI, blockchain and talent recruitment organizations using three front companies, according to Microsoft.
Since late October 2026, Onyx Sleet and Pearl Sleet have leveraged the open-source agentic AI penetration-testing framework ARTEX in operations, Microsoft said.
Onyx Sleet is known domestically as Andariel, and Onyx Sleet and Pearl Sleet are under North Korea’s Reconnaissance General Bureau, ChosunBiz reported.
Microsoft assessed that North Korean AI-enabled operations remain human-directed and have not become fully autonomous.





