In recent months, Google Threat Intelligence Group (GTIG) has identified a rise in threat actors leveraging AI for information gathering, creating hyper-realistic phishing scams, and developing sophisticated malware. Although there have been no direct attacks on frontier models or generative AI products by advanced persistent threat (APT) actors, GTIG has mitigated numerous model extraction attacks— a significant form of corporate espionage—affecting organizations globally. This highlights a looming threat for other businesses with AI models. Our newly released report outlines these findings and details proactive measures we’ve implemented, including disabling malicious accounts and enhancing security controls for our Gemini models. By addressing these vulnerabilities, we aim to protect our infrastructure and customers from potential misuse. For comprehensive insights, read the full report on the Google Cloud Threat Intelligence blog.
This summary integrates critical SEO terms while maintaining a concise overview of GTIG’s findings and actions.
Source link
