We're sunsetting PodQuest on 2025-07-28. Thank you for your support!
Export Podcast Subscriptions
cover of episode EP 560: Inside Multi-Agentic AI: 3 Critical Risks and How to Navigate Them

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

2025/7/3
logo of podcast Everyday AI Podcast – An AI and ChatGPT Podcast

Everyday AI Podcast – An AI and ChatGPT Podcast
AI Deep Dive AI Chapters Transcript
People
S
Sarah Bird
Topics
Sarah Bird: 随着生成式AI和代理AI的快速发展，负责任的AI也随之发生巨大变化。过去，我们只需要关注人与AI聊天机器人之间的互动，但现在我们需要考虑代理AI之间的互动，以及由此产生的伦理、治理和责任问题。我的团队会关注AI系统中出现的新风险，并构建工具和技术来测试和缓解这些风险，使每个人都能成功地使用AI。虽然确保AI符合我们的原则并以负责任的方式构建和使用AI的目标没有变，但人们对AI重要性的认识和参与度大大提高。代理AI之所以发展迅速，是因为它能真正实现技术的承诺，能够帮助人们完成任务，而不仅仅是聊天。但与此同时，代理AI能够代表用户执行任务，因此出错的可能性更大，影响也可能更大。由于代理AI可以在没有人工干预的情况下工作更长时间，因此我们失去了人工监督这一重要缓解措施。我们需要调整系统来管理、保护和管理这些新的实体，就像我们今天管理用户、应用程序和设备一样，同时也要解决代理AI带来的新风险，例如偏离任务或意外泄露敏感数据。

Deep Dive

Chapters
The field of Responsible AI has drastically changed with the advent of generative AI and agentic AI. It used to be simpler to understand the guardrails needed for AI, but now, with agents making decisions, the ethical, governance, and responsibility aspects have become significantly more complex. The increased awareness and engagement from businesses regarding responsible AI is a positive development.
  • The rapid evolution of generative AI and agentic AI has made responsible AI significantly more complex.
  • Businesses are increasingly prioritizing responsible AI from the outset of their AI projects.
  • The maturity and understanding of responsible AI have grown significantly in recent years.

Shownotes Transcript

Multi-agentic AI is rewriting the future of work.... but are we racing ahead without checking for warning signs?Microsoft’s new agent systems can split up work, make choices, and act on their own. The possibilities? Massive.But it's not without risks, which is why you NEED to listen to Sarah Bird. She's the Chief Product Officer of Responsible AI at Microsoft and is constantly building out safer agentic AI. So what’s really at stake when AIs start making decisions together?And how do you actually stay in control?We’re pulling back the curtain on the 3 critical risks of multi-agentic AI and unveiling the playbook to navigate them safely.

Newsletter: Sign up for our free daily newsletter)**More on this Episode: **Episode Page)**Join the discussion: **Have a question? Join the convo here.)Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineup)Website: YourEverydayAI.com)Email The Show: [email protected])**Connect with Jordan on **LinkedIn)

Topics Covered in This Episode:

  • Responsible AI: Evolution and Challenges
  • Agentic AI's Ethical Implications
  • Multi-Agentic AI Responsibility Shift
  • Microsoft’s AI Governance Strategies
  • Testing Multi-Agentic Risks and Patterns
  • Agentic AI: Future Workforce Skills
  • Observability in Multi-Agentic Systems
  • Three Risk Categories in AI Implementation

**Timestamps:**00:00 Evolving Challenges in Responsible AI

05:50 Agent Technology: Benefits and Risks

09:27 Complex System Governance and Observability

12:26 AI Monitoring and Human Intervention

15:14 Essential Testing for Trust Building

19:43 Securing AI Agents with Entra

22:06 Exploring Human-AI Interface Innovation

26:06 AI Workforce Integration Challenges

28:22 AI's Transformative Impact on Jobs

**Keywords:**Agentic AI, multi agentic AI, responsible AI, generative AI, Microsoft Build conference, AI governance, AI ethics, AI systems, AI risk, AI mitigation, AI tools, human in the loop, Foundry observability, AI testing, system security, AI monitoring, user intent, AI capability, prompt injection, Copilot, AI orchestration, AI deployment, system governance, Entra agent ID, AI education, AI upskilling, AI workforce integration, systemic risk, AI misuse, AI malfunctions, AI systemic risk, AI-powered solutions, AI development, AI innovation, AI technology, AI security measures.

Send Everyday AI and Jordan a text message. (We can't reply back unless you leave contact info) )

Ready for ROI on GenAI? Go to youreverydayai.com/partner )