The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) How Microsoft Scales Testing and Safety for Generative AI with Sarah Bird - #691
9 snips
Jul 1, 2024 Join Sarah Bird, Chief Product Officer of Responsible AI at Microsoft, as she dives into the essential realms of generative AI testing and safety. Explore the challenges of AI hallucinations and the importance of balancing fairness with security. Hear insights from Microsoft's past failures like Tay and Bing Chat, stressing the need for adaptive testing and human oversight. Sarah also discusses innovative methods like automated safety testing and red teaming, emphasizing a robust governance framework for evolving AI technologies.
AI Snips
Chapters
Transcript
Episode notes
Generative AI Risk Categories
- Generative AI risks involve undesirable content and problematic model capabilities.
- Focus on both content types (e.g., harmful code) and model behaviors (e.g., hallucinations).
Learning from Tay and Bing Chat
- Microsoft learned from Tay's failure and adapted after Bing Chat's unexpected behaviors.
- The Bing Chat incident highlighted the need for extensive testing beyond initial expectations.
Layered Defenses
- Implement layered defenses for AI systems, combining various safety technologies.
- Each technology has strengths and weaknesses, creating a comprehensive approach like "Swiss cheese."