The AI Study Guide: #MarchResponsibly with AI

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

Hi! It's me, Natalie, your Azure AI Skilling Guru. Want to learn something about Azure AI or ML? We've got something for that!Hi! It's me, Natalie, your Azure AI Skilling Guru. Want to learn something about Azure AI or ML? We've got something for that!

Welcome to the March edition of the Azure AI Study Guide. Every month I’ll bring you the best and newest tools when it comes to skilling up on Azure AI. This month, I’m partnering with my friend and Microsoft Cloud Advocate Ruth Yakubu for a #MarchResponsibly campaign - educating AI and Machine Learning enthusiasts about responsible AI.  


Join us for this month-long celebration of data science, machine learning and artificial intelligence. You will see a series of blog posts that will cover the tools, technologies, and social impact, of responsible AI. Let’s #MarchResponsibly together. As part of this blog series, we’ve curated the best of Microsoft’s learning resources so you can build and test fair AI solutions.


I’ve brought in a mix of videos, learning modules and tutorials that will give you a great foundation for deploying AI solutions responsibly.


Let’s get started!


What is responsible AI and why is it important?

Responsible AI is an approach to developing, assessing, and deploying AI systems in a safe, trustworthy, and ethical way. This means putting values such as fairness, reliability, and transparency in the middle of our system design decisions. Responsible AI technology ensures that your AI systems are designed, developed, and deployed in a responsible and fair manner. If you’re interested in learning more about Microsoft’s responsible AI practices, see the home page here.


The Basics

Before we dive into technical skills, I suggest starting with these overviews:

Embrace responsible AI principles and practices module: In this quick module (50 minutes) you’ll get an overview that’s fit for anyone, whether you’re a developer or a business leader (or both!).


Fundamentals of Responsible Generative AI: This module is a hot item – in under an hour you’ll get prepared to deploy and operate a generative AI solution responsibly.


I also really enjoyed this three-part blog series that dives into the basics and the tools you need to get started:

Responsible AI in action, Part 1: Get started

Responsible AI in action,  Part 2: Complete an impact assessment

Responsible AI in action, Part 3: Tools to help


The HAX Toolkit is an awesome resource for teams building user-facing AI products. It’s all about creating human-centered AI experiences and testing your protypes early on. 


Responsible AI tools to debug models on Azure Machine Learning


Module: In this FREE 60 minutes hands on lab module, you’ll learn how to use the Responsible AI Dashboard (one of the tools reviewed above) with Azure Machine Learning to debug models and make data-driven decisions.  Here’s a quick video to give you an idea of the capabilities and functionality of the dashboard. 


GitHub Repository: This repo feeds into the module above and includes even more learning opportunities to explore multiple datasets and debug them.


Responsible AI Mitigation and Tracker. This toolkit helps AI Engineers explore mitigation steps needed to interactively track and compare mitigation experiments. It enables data scientists to see where the model has improved and whether there are variations in performance for different data cohorts (subgroups).


Responsible AI Toolbox: Finally, a suite of tools just for AI and ML practitioners! In this toolbox you’ll get a review of 4 different tools (including the Responsible AI Dashboard) for model assessment and decision making resulting in a customized, end to end responsible AI experience. For a nice review plus a knowledge check, I recommend the toolbox landing page.


SmartNoise SDK Toolkit: This toolkit is new to me and I’m so glad my colleagues shared it.  This differential privacy toolkit is for analytics and machine learning. It injects noise into data to prevent disclosure of sensitive information. How cool is that? Click “Get Started” on the home page to access the GitHub page.


Responsible AI and Azure AI Studio

First, review the fundamentals of responsible generative AI module linked above.


I also recommend the Introduction to Azure AI Studio module (less than 1 hour) to get oriented in Azure AI Studio since we are going to be focusing here a lot.


Next, let’s check out this awesome blog series by our own Sarah Bird. She gives an amazing overview of Azure AI Content Safety and how it can be used within the new Azure AI Studio.

Sarah Bird also joined Seth Juarez on his AI Show – check out their quick 25 minute video visual overview and demo of Azure AI Content Safety. 


The next video in the series dives into the how to use Azure AI Studio to evaluate your app’s performance.


This tutorial will give you a guided experience in trying your hand at the prompt flow Content Safety tool in Azure AI Studio.


You can also always check out my friend Ruth’s Microsoft Learn collection for more tips and tricks on responsible AI.


And as always, keep an eye on the Microsoft Learn AI Hub for the latest learning opportunities for teams and individuals.

Leave a Reply

Your email address will not be published. Required fields are marked *


This site uses Akismet to reduce spam. Learn how your comment data is processed.