IBM introduces a new security framework, the IBM Framework for Securing Generative AI, to address the risks posed by gen AI. The framework focuses on protecting gen AI workflows and provides guidance on security threats and defensive approaches. Core tenets include securing the data, model, and usage, with secure infrastructure and AI governance. Unique risks in gen AI include data poisoning, bias and data diversity, data drift, data privacy, and prompt injection.
Table of contents
What’s different about gen AI security?MLSecOps, Machine Learning Detection and Response and the new AI security landscapeSort: