@techreport{vorvoreanu2023responsible,
author = {Vorvoreanu, Mihaela and Heger, Amy and Passi, Samir and Dhanorkar, Shipi and Kahn, Zoe and Wang, Ruotong},
title = {Responsible AI Maturity Model},
institution = {Microsoft},
year = {2023},
month = {May},
abstract = {The Responsible AI Maturity Model (RAI MM) is a framework to help organizations plan how to achieve RAI maturity.

Organizations can use the RAI MM to:

 	Plan their RAI strategy & programs (policy, processes, infrastructure)
 	Help their AI product teams understand the sociotechnical nature of RAI and engage in the cross-discipline collaborations that are key to RAI
 	Check on their RAI progress

The RAI MM is not just a paper to read - it is a tool to use in strategic planning workshops. Contact the team if you need help with running a workshop - or use the materials below:

Download executive summary & FAQ

Download the RAI MM

Download the RAI MM workshop facilitator deck

 

The RAI MM contains 24 empirically derived dimensions that are key to an organization’s RAI maturity. The dimensions and their levels are based on interviews and focus groups with over 90 RAI specialists (e.g., RAI champs, MSR experts) and AI practitioners (e.g., user experience (UX) researchers, UX designers, data scientists). Each dimension has five levels, going from low (Level 1: Latent) to high (Level 5: Leading) maturity. The dimensions are organized into three main categories:

 	Organizational Foundations
 	Team Approach
 	RAI Practice

Engage key stakeholders from different business units in a RAI strategic planning workshop. Follow these steps:

 	Together, select a few RAI maturity dimension you want to focus in a given time frame (e.g., the next 6 months).
 	For each selected dimension, decide what maturity level you want to aim for.
 	Make a list of action items needed to reach the desired maturity level for each selected dimension.

We welcome your feedback!},
url = {http://approjects.co.za/?big=en-us/research/publication/responsible-ai-maturity-model/},
number = {MSR-TR-2023-26},
}