Catalogue of Tools & Metrics for Trustworthy AI

These tools and metrics are designed to help AI actors develop and use trustworthy AI systems and applications that respect human rights and are fair, transparent, explainable, robust, secure and safe.

Microsoft's Responsible AI Toolbox



Microsoft's Responsible AI Toolbox is an open-source suite of tools for assessing, debugging and improving machine learning models. It was developed by Microsoft. It combines interactive dashboards and code libraries that help developers and other stakeholders understand how AI systems behave, identify problems and make informed, data-driven decisions. It supports models built on tabular data, text and images.

The central component is the Responsible AI dashboard, which brings several assessment tools together in one interface. Error analysis identifies groups of data for which a model performs worse than average, such as specific demographic groups or rare input conditions. Fairness assessment, powered by Fairlearn, shows which groups of people may be disproportionately harmed by a model. Model interpretability, powered by InterpretML, explains a model's overall behaviour and the reasons behind individual predictions. Counterfactual analysis shows the smallest change to an input that would lead to a different outcome, such as the higher income that would have led to a loan approval.

Causal analysis estimates the real-world effect of possible interventions, helping decision-makers answer "what if" questions. Data balance analysis shows whether some groups are underrepresented or favoured in the data. Users can combine these components into custom workflows, for example moving from error analysis to data exploration to diagnose the source of a problem. Related repositories add libraries for mitigating identified issues, tracking model experiments and measuring gender bias in text datasets. The toolbox is written in Python and TypeScript and released under the MIT licence.

Auto-discovered on 2026-02-11 by OECD Catalogue Automation

About the tool


Developing organisation(s):


Tool type(s):




Country/Territory of origin:



Type of approach:







Risk management stage(s):


Programming languages:



Github stars:

  • 1839

Github forks:

  • 499

Modify this tool

Use Cases

There is no use cases for this tool yet.

Would you like to submit a use case for this tool?

If you have used this tool, we would love to know more about your experience.

Add use case
Partnership on AI

Disclaimer: The tools and metrics featured herein are solely those of the originating authors and are not vetted or endorsed by the OECD or its member countries. The Organisation cannot be held responsible for possible issues resulting from the posting of links to third parties' tools and metrics on this catalogue. More on the methodology can be found at https://oecd.ai/catalogue/faq.