top of page


Dwain.B
2 Feb 2025
Adversarial Misuse of Generative AI
Google has just released a comprehensive document detailing how government-backed threat actors are attempting to misuse its AI model, Gemini. The report covers AI-driven cyber threats, information operations, and efforts to bypass AI safety measures.
🔎 Key insights:
Government-backed hackers from Iran, China, North Korea, and Russia have tested Gemini for coding, reconnaissance, and content manipulation.
AI jailbreak attempts were largely unsuccessful, but actors found ways to enhance operations.
The report highlights Google's commitment to securing AI and countering misuse.
Download here:
bottom of page