Microsoft Azure Prompt Shield Guardrail Evasion

Affected Vendor(s)

Affected Product(s)

Summary

Azure Prompt Shield is a jailbreak and prompt injection classifier created by Microsoft. Prompt Shield can be enabled by developers using Azure OpenAI to protect their AI applications from these threats. We successfully demonstrated how an attacker can fully evade, or greatly degrade, classification accuracy of the classifier, enabling prompt injections and jailbreaks to pass through filters and subsequently to the protected AI application.

Timeline

Discovered on
Disclosed to Vendor on
June 6, 2024
Published on
June 24, 2024

Credit

Blog Post

References

Learn how Mindgard can help you navigate AI Security

Take the first step towards securing your AI. Book a demo now and we'll reach out to you.