Comprehensive AI Safety and Security with defense in depth for Enterprises

This post has been republished via RSS; it originally appeared at: Microsoft Tech Community - Latest Blogs - .

Azure AI Content Safety APIs

Azure AI Content Safety is a new service that helps detect hateful, violence, sexual, and self-harm content in images and text, and assign severity scores, allowing businesses to limit and prioritize what content moderators need to review. Unlike most solutions used today, Azure AI Content Safety can handle nuance and context, reducing the number of false positives and easing the load on human content moderator teams.

Prompt Shields (preview)	Identify and block direct and indirect prompt injection attacks before they impact your model, scans text for the risk of a User input attack on a Large Language Model. Quickstart
Groundedness detection (preview)	It detect model “hallucinations” so you can block or highlight ungrounded responses, detects whether the text responses of large language models (LLMs) are grounded in the source materials provided by the users. Quickstart
Protected material text detection (preview)	Blocks copyrighted or known content like song lyrics, scans AI-generated text for known text content (for example, song lyrics, articles, recipes, selected web content). Quickstart
Custom categories (rapid) API (preview)	Create and deploy your own content filters, lets you define emerging harmful content patterns and scan text and images for matches. How-to guide
Analyze text API	Scans text for sexual content, violence, hate, and self harm with multi-severity levels.
Analyze image API	Scans images for sexual content, violence, hate, and self harm with multi-severity levels.