OpenAI Disrupts Hacking Groups Using ChatGPT for Cyberattacks

State-Linked Actors Weaponize AI

OpenAI on Oct. 8, 2025, announced that it had disrupted and banned three state-linked hacking groups from Russia, North Korea, and China that were systematically misusing ChatGPT for malicious purposes. The disclosure confirms that sophisticated threat actors have integrated large language models (LLMs) into their standard operational toolkit to accelerate their existing workflows.

OpenAI stated the groups used its models as a productivity tool for tasks including creating modular code for malware, developing command-and-control infrastructure, and crafting multilingual phishing campaigns. The company said it had banned the accounts and shared its findings with security partners.