Rebekka Weiss & Tobi Müller

Responsible AI @ Microsoft - Governance, Standards, Learnings

How does Microsoft govern AI as both a powerful tool and a potential weapon? See the framework they use to move from abstract principles to concrete practice.

Responsible AI @ Microsoft - Governance, Standards, Learnings
#1about 2 minutes

Responsible AI is more than just legal compliance

Regulation shapes AI governance, but true responsibility also involves code, shared learnings, and operational standards.

#2about 3 minutes

Establishing foundational principles for AI development

Microsoft's AI strategy began by proactively addressing risks and establishing core principles like fairness, privacy, and security.

#3about 3 minutes

Addressing new risk vectors in generative AI

Generative AI introduces unique risks like jailbreaks, prompt injection, and harmful content that require continuous mitigation efforts.

#4about 2 minutes

Using the NIST framework to structure AI risk management

The NIST AI framework provides a standardized approach to map, measure, manage, and govern AI risks effectively.

#5about 4 minutes

Building a diverse and collaborative governance model

Microsoft's hub-and-spokes model for responsible AI relies on diverse teams and thorough documentation to share learnings and prevent duplicated efforts.

#6about 3 minutes

Structuring the office of responsible AI for impact

The Office of Responsible AI is organized into three key pillars—engineering, policy, and research—to address the multifaceted nature of AI governance.

#7about 3 minutes

Securing systems with red teaming and data governance

Proactive security measures like large-scale red teaming and robust data governance are essential for protecting AI systems and user privacy.

#8about 2 minutes

Scaling risk identification from manual to automated testing

The risk assessment process begins with manual human testing to understand user behavior, which is then scaled using automated tools like PyRIT.

#9about 4 minutes

Prioritizing human oversight with layered safety mitigations

The "Copilot, not Autopilot" philosophy emphasizes irreplaceable human review and a multi-layered mitigation strategy for safe human-AI interaction.

#10about 2 minutes

Ensuring product readiness and shaping global AI standards

A rigorous pre-launch assessment ensures product safety, while advocacy for global standards aims to make AI beneficial for everyone.

Related jobs
Jobs that call for the skills explored in this talk.

Matching moments

Implementing responsible AI with governance and sandboxing
10:55 MIN

Implementing responsible AI with governance and sandboxing

Beyond the Hype: Real-World AI Strategies Panel

The current state of responsible AI in the private sector
05:14 MIN

The current state of responsible AI in the private sector

Responsible AI in Practice: Real-World Examples and Challenges

Implementing responsible AI and data privacy in enterprises
14:46 MIN

Implementing responsible AI and data privacy in enterprises

Architecting the Future: Leveraging AI, Cloud, and Data for Business Success

A practical framework for building responsible AI
19:42 MIN

A practical framework for building responsible AI

A walkthrough on Responsible AI Frameworks and Case Studies

Defining the meaning of responsible AI in business
01:19 MIN

Defining the meaning of responsible AI in business

Responsible AI in Practice: Real-World Examples and Challenges

Defining the core principles of responsible AI
06:12 MIN

Defining the core principles of responsible AI

Rethinking Recruiting: What you didn’t know about Responsible AI

Making the business case for responsible AI
23:41 MIN

Making the business case for responsible AI

Responsible AI in Practice: Real-World Examples and Challenges

Managing the fear, accountability, and risks of AI
14:10 MIN

Managing the fear, accountability, and risks of AI

Collaborative Intelligence: The Human & AI Partnership

Featured Partners

Related Videos

See all videos
Responsible AI in Practice: Real-World Examples and Challenges29:35

Responsible AI in Practice: Real-World Examples and Challenges

Steffen Bosse, Mina Saidze, Ray Eitel-Porter & Björn Bringmann

Trust by Design: Creating Responsible AI-Powered Services25:51

Trust by Design: Creating Responsible AI-Powered Services

Dr. Marc Fuchs, Christoph Bräunlein, Eva Stepkes & Niklas Harzheim

Rethinking Recruiting: What you didn’t know about Responsible AI28:07

Rethinking Recruiting: What you didn’t know about Responsible AI

Vincent Slot & Jaap Kersten

A walkthrough on Responsible AI Frameworks and Case Studies28:05

A walkthrough on Responsible AI Frameworks and Case Studies

Toju Duke

Inside the AI Revolution: How Microsoft is Empowering the World to Achieve More1:09:49

Inside the AI Revolution: How Microsoft is Empowering the World to Achieve More

Simi Olabisi

Kill Switch or Moral Compass: Who Programs AI’s Conscience?28:54

Kill Switch or Moral Compass: Who Programs AI’s Conscience?

Torsten Stiller

GenAI Security: Navigating the Unseen Iceberg24:33

GenAI Security: Navigating the Unseen Iceberg

Maish Saidel-Keesing

AI in Action: Real Use Cases with Real Impact - Hanna Hennig, Michael Ameling, Tobias Regenfuss29:00

AI in Action: Real Use Cases with Real Impact - Hanna Hennig, Michael Ameling, Tobias Regenfuss

Hanna Hennig, Michael Ameling & Tobias Regenfuss and Mike Butcher

Related Articles

View all articles
CH
Chris Heilmann
Exploring AI: Opportunities and Risks for Developers
In today's rapidly evolving tech landscape, the integration of Artificial Intelligence (AI) in development presents both exciting opportunities and notable risks. This dynamic was the focus of a recent panel discussion featuring industry experts Kent...
Exploring AI: Opportunities and Risks for Developers
CH
Chris Heilmann
With AIs wide open - WeAreDevelopers at All Things Open 2025
Last week our VP of Developer Relations, Chris Heilmann, flew to Raleigh, North Carolina to present at All Things Open . An excellent event he had spoken at a few times in the past and this being the “Lucky 13” edition, he didn’t hesitate to come and...
With AIs wide open - WeAreDevelopers at All Things Open 2025
CH
Chris Heilmann
WWC24 Talk - Scott Hanselman - AI: Superhero or Supervillain?
Join Scott Hanselman at WWC24 to explore AI's role as a superhero or supervillain. Scott shares his 32 years of experience in software engineering, discusses AI myths, ethical dilemmas, and tech advancements. Engage with his live demos and insights o...
WWC24 Talk - Scott Hanselman - AI: Superhero or Supervillain?

From learning to earning

Jobs that call for the skills explored in this talk.