Google says hackers abuse Gemini AI to empower their attacks

Google's Threat Intelligence Group reports that state-sponsored APTs from over 20 countries — particularly Iran, China, North Korea, and Russia — have been experimenting with Google's Gemini LLM to assist with reconnaissance, vulnerability research, scripting, post-compromise activities, and content creation; usage has primarily improved attacker productivity rather than produced novel AI-enabled attacks, but it shortens preparation time and raises broader abuse risks as less-protected models are easy to exploit.