OpenAI confirms threat actors use ChatGPT to write malware

OpenAI confirmed multiple Chinese and Iranian-aligned threat actors abused ChatGPT to enhance offensive cyber operations — including spear-phishing that delivered SugarGh0st RAT, development and debugging of Android infostealers and C2 infrastructure, LLM-assisted reconnaissance and vulnerability research, code obfuscation, and post-compromise credential harvesting — demonstrating generative AI can materially increase attacker efficiency; accounts were banned and IOCs shared with partners.