"Wiz Discovers Flaws in GenAI Models Enabling Customer Data Theft"
"Wiz Discovers Flaws in GenAI Models Enabling Customer Data Theft"
Wiz, a cloud security provider, discovered two critical architecture flaws in generative Artificial Intelligence (AI) models uploaded to Hugging Face, the leading AI model and application-sharing platform. Wiz Research described the two flaws and the potential risk they pose to AI-as-a-service providers. The risks include shared inference infrastructure takeover and shared Continuous Integration and Continuous Deployment (CI/CD) takeover. This article continues to discuss how attackers could exploit the AI infrastructure risks.