Microsoft Responsible AI
Tools and practices
Evaluate, understand, and make informed decisions about your AI systems.
Featured tools and processes
Discover tools that can help your organization map, measure, and manage AI risks throughout the development cycle to reduce the risk of harm.
Microsoft Responsible AI Standard
Explore Microsoft internal guidance on how to design, build, and test AI systems.
AI Impact Assessment Guide
Explore internal guidance from Microsoft for assessing the impact of AI.
AI Impact Assessment Template
Get the Microsoft template for assessing the impact of AI.
Human-AI Experience Toolkit
Use this toolkit to plan and design human-centered AI systems.
Responsible AI Toolbox
Explore open-source interfaces and libraries that help developers better understand and monitor AI systems.
Azure AI Content Safety
Automatically detect and filter unsafe content in generative AI prompts and outputs for your application.
Microsoft Purview
Safeguard and manage compliance of data in AI tools and systems.
AI Impact Assessment Guide
Explore Microsoft internal guidance for assessing the impact of AI.
AI Impact Assessment Template
Get the Microsoft template for assessing the impact of AI.
AI Red Teaming
Learn how to safeguard your organization's AI with guidance and best practices from the industry leading Microsoft AI Red Team.
Threat modeling
Gain perspective and resources to adopt threat modeling as part of the software development lifecycle (SDLC).
Accelerate Red Teaming with PyRIT
Learn how the open automation framework empowers red teams to uncover risks in generative AI systems.
Failure Modes in Machine Learning
Identify threats, attacks, and vulnerabilities—and plan for countermeasures—using this framework.
Inclusive Design
Explore Microsoft’s inclusive design methodology which draws on a wide range of diverse perspectives.
Responsible AI dashboard
Assess and improve model fairness, accuracy, and explainability with the Responsible AI dashboard.
Fairlearn
Assess your AI system's fairness, then mitigate any observed bias, by installing this open-source package.
InterpretML
Understand ML models, including glass-box models and blackbox systems, with help from this open-source toolkit.
Error Analysis
Identify and diagnose ML model errors with this responsible AI toolkit for improving model accuracy.
Evaluate GenAI apps
Assess generative AI apps for quality and safety using human review and AI-assisted metrics in Azure AI Foundry.
EconML
Estimate causal effects or responses from observational data using ML techniques with this Python package.
Counterfit
Assess the security of your ML systems using a generic automation layer.
Microsoft Purview
Safeguard and manage the compliance of all your data for AI tools and systems.
Confidential computing
Increase data privacy and security surrounding business and consumer data by protecting data in use for AI.
Microsoft Defender for Cloud
Proactively mitigate risks and identify threats to AI apps, from code to cloud, with comprehensive security.
Microsoft SEAL
Perform computations directly on encrypted data with this open-source homomorphic encryption library.
Presidio
Protect sensitive data with customizable, context-aware de-identification features for text and images.
Azure AI
Responsibly build intelligent apps at enterprise scale with the Azure AI portfolio.
Azure Policy
Achieve organization-wide cloud compliance by setting guardrails and creating policies to govern your resources.
Responsible AI Toolbox
Access a suite of open-source tools designed to help you assess, develop, and deploy AI responsibly.
Business resources
AI for Leaders
Discover how to drive business impact by creating effective AI strategies.
CUSTOMER STORIES
See responsible AI innovations across industries
AI learning
Build your AI skills
Deepen your understanding of how to use and develop AI responsibly.
Explore products and solutions
Azure AI Foundry
Design, customize, and monitor AI apps and agents at scale.
Microsoft Security Copilot
Protect your organization with an AI-powered assistant for security and IT operations.
