Everyday AI Podcast – An AI and ChatGPT Podcast

EP 560: Inside Multi-Agentic AI: 3 Critical Risks and How to Navigate Them

71 snips
Jul 3, 2025
Guest
Sarah Bird
Sarah Bird, Chief Product Officer of Responsible AI at Microsoft, shares her insights on the future of multi-agentic AI. She highlights the incredible potential of these systems to transform work by acting autonomously but warns of the critical risks involved. The conversation explores ethical implications, the shift in AI responsibility, and the essential need for governance strategies. Bird emphasizes the importance of proactive testing and developing skills for effective human-AI collaboration to navigate this emerging landscape safely.
Ask episode
AI Snips
Chapters
Transcript
Episode notes
INSIGHT

Agentic AI Changes Control Dynamics

  • Agentic AI represents a major leap where systems do tasks autonomously for users, fulfilling the promise of ease and automation.
  • This increases surface area for risks as humans are less in direct control, requiring new approaches to oversight and governance.
INSIGHT

Multi-Agent AI as Modular Task System

  • Current multi-agent AI often involves breaking a large task into subtasks handled by specialized agents.
  • This modular approach enables more precise testing and safer governance of each component.
ADVICE

Govern Each Agent Separately

  • Govern each agent individually within a multi-agent system to maintain visibility and security.
  • Utilize monitoring tools like Foundry observability to detect off-task behavior and tool misuse early.
Get the Snipd Podcast app to discover more snips from this episode
Get the app