Know What's Happening in Business

Google Gemini Weaponized in State-Sponsored Attacks

Google’s Gemini large‑language model is being weaponized by multiple state‑sponsored threat actors, according to the Google Threat Intelligence Group. North Korean UNC2970 and several Chinese groups such as Mustang Panda, Judgment Panda, APT41 and UNC795 are using Gemini for rapid OSINT synthesis, vulnerability analysis, exploit debugging, and web‑shell creation. Iranian APT42 and UNC6418 exploit the AI assistant for social‑engineering and target intelligence, while the new HONESTCUE malware calls the Gemini API to generate code for its second‑stage payload. The report highlights a growing trend of AI‑driven cyber‑offensives across diverse geopolitical actors.