Adversarial Misuse of Generative AI
2025-01-29 • Google •
https://cloud.google.com/blog/topics/threat-intelligence/adversarial-misuse-generative-ai
Google Threat Intelligence Group examined how government-backed APT and information-operations actors attempted to use Gemini for operational support. The source says actors used the tool mainly for research, troubleshooting, content generation, localization, infrastructure research, reconnaissance, vulnerability research, payload development, and scripting assistance. GTIG did not observe novel AI-enabled attack techniques or successful prompt attacks, and it says Gemini did not produce malware or other content that could plausibly enable a successful malicious campaign in the reviewed activity.