Human in the Loop

What happens when Responsible AI is left out of the conversation? In this episode of Human in the Loop, Sean Astrakhan joins us to discuss what Responsible AI means to him, whether it truly matters, and how it will influence the future of work for all of us.
What you will learn:

• Why AI success depends on accuracy, cost-effectiveness, and security (ACS).
• How to think about responsible and trustworthy AI beyond just ethics.
• Why security, governance, and human oversight remain critical in AI adoption.
• How AI is changing jobs and why continuous upskilling is essential.
• The risks of banning AI tools and driving employees toward shadow AI.
• Why understanding people and business problems matters more than technical skills alone.
• How finding your passion and surrounding yourself with ambitious people can accelerate your career.

As AI agents become more capable, measuring and mitigating risk is more important than ever. In this episode of Human in the Loop, our guest Mehrnoosh Sameki discusses the shift from evaluation to mitigation, the new open-source measurement framework ASSERT, and the role of agent control systems in building safer and more trustworthy agents.

What You Will Learn:

• Understand the unique risks posed by autonomous AI agents.
• Learn why AI agent testing requires more than standard safety checks.
• Discover how ASSERT helps evaluate AI systems against custom policies and regulations.
• Explore how context-aware testing improves AI reliability and safety.
• Learn why sandboxing is essential for safe adversarial testing.
• See how Agent Control Spec (ACS) enables scalable AI governance and controls.
• Understand the journey from AI evaluation to mitigation and continuous monitoring.

Join Daniel Laskewitz, Principal Cloud Advocate at Microsoft, for a conversation exploring the intersection of Model Context Protocol (MCP), Copilot Studio, and Responsible AI (RAI). We discuss how MCP is shaping the next generation of AI integrations, how organizations can build powerful agents with Copilot Studio, and why responsible AI principles must remain at the center of agent design, deployment, and governance. A practical discussion for developers, makers, and AI leaders looking to build intelligent systems that are both capable and trustworthy.

What You Will Learn:

• The latest innovations coming to Copilot Studio.
• How MCP is transforming AI integrations.
• Why less instruction can sometimes create better AI agents.
• What multi-agent architectures look like in practice.
• When to use Copilot, Cowork, Scout, or Copilot Studio.
• The opportunities and risks of AI skills marketplaces.
• Why responsible AI still requires humans in the loop.

This episode explores data poisoning and its growing impact on AI systems, from model backdoors to agent memory risk. Ioana and Chris chat with Microsoft's Giorgio Severi about how adversaries manipulate data, why these attacks are hard to detect, and what it takes to build layered defenses that keep AI systems reliable, safe, and trustworthy.

What You Will Learn:

• Understand what AI red teaming is and why it’s critical for safe AI deployment
• Learn how data and model poisoning can subtly influence AI behavior over time
• Explore why AI systems can fail silently (e.g., backdoors and hidden triggers)
• Discover the importance of layered security (“defense in depth”) in AI systems
• Gain insight into new risks in AI agents, especially around memory and persistence
• Get practical guidance on how to design and test more trustworthy AI systems

Publications:
[2602.03085] The Trigger in the Haystack: Extracting and Reconstructing LLM Backdoor Triggers
[1708.06733] BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chain

#EnterpriseAI #AISafety #TrustworthyAI #AIGovernance #AgenticAI #AIResearch

For this episode of Human in the loop, we talked with Dux Raymond Sy (Chief Brand Officer, AvePoint) to explore why a foundation of trusted data is crucial in the AI era. We chatted about how to avoid data traps, how to prepare for a future full of AI assistants, and turn AI hype into real results without compromising trust.

What You Will Learn:

• Why trustworthy data is the foundation of effective and safe AI, especially as systems act autonomously.
• What “good data” really means; data you can trust, govern, and automate decisions with.
• How AI agents and multi‑agent systems amplify risk when data quality and controls are weak.
• Why visibility, ownership, and lifecycle management of AI agents are now critical.
• How to reframe governance as guardrails that enable people to succeed, not slow them down.
• Why human‑in‑the‑loop, monitoring, and kill switches are essential for autonomous AI.
• How different regions are approaching AI adoption and education, and what others can learn from them.

  #HumanInTheLoop #TrustworthyAI #AIGovernance #ResponsibleAI #AgenticAI

In this episode of Human in the Loop, we sit down with Meredith from Applied Information Sciences to explore what it truly takes to bring AI into organizations responsibly. From securing sensitive data to navigating shadow AI, she shares practical guidance shaped by years of working with highly regulated industries. Together, we unpack the evolving role of AI policies, the importance of human‑in‑the‑loop safeguards, and how emerging agentic systems are reshaping expectations for security and governance. It’s a grounded, insightful conversation for anyone looking to innovate with AI while keeping trust and safety at the center.

What You Will Learn:
• Trustworthy AI is crucial for responsible innovation
• Data security and governance are often underestimated risks.
• Zero Trust principles are essential in AI security.
• AI amplifies existing cybersecurity risks.
• Organizations must address shadow AI usage.
• Policies around AI usage are lacking in many organizations.
• Everyone in the organization shares responsibility for AI governance.
• Human oversight is necessary in AI decision-making.
• Experimentation with AI tools can lead to valuable insights.
• Curiosity and willingness to try new AI tools are vital.

In this episode of Human in the Loop, we sit down with Meeri Haataja, CEO and co-founder of Saidot, to explore what responsible AI governance really looks like inside organizations. From navigating regulatory frameworks to embedding transparency and accountability into AI development, Meeri shares practical insights for businesses at every stage of their AI journey. Whether you're scaling a startup or steering enterprise innovation, her advice on aligning technical ambition with ethical clarity is essential listening.

What you will learn:
• AI governance is crucial for building trust in AI systems.
• Responsible AI involves understanding and managing risks.
• The EU AI Act is gaining awareness but still needs more attention
• Organizations must recognize their responsibilities as AI deployers.
• Procurement processes are becoming more strategic in AI governance.
• Regulations provide a solid foundation for responsible AI practices.
• AI governance can differentiate businesses in the market.
• Experimentation with AI agents is essential for learning.
• Scaling AI governance is a significant challenge for organizations.
• Education and awareness are key to effective AI governance.

In this very first episode, hosts Chris Huntingford and Ioana Tanase, talk with Microsoft's Altrese Hawkins-Gopie about what responsible AI really means, why it matters for everyone (not just techies), and how we can all play a part in building Trustworthy AI.

What you will learn:
• What “responsible AI” is (with real-world examples)
• Why community feedback and inclusion are essential
• How users can help shape better AI (even if you’re not a developer)
• The importance of fact-checking and not blindly trusting AI
• Practical steps for organizations to build responsible AI
• How AI is changing the future of work and the skills you’ll need

In this episode of Human in the Loop, Christina Mallon joins us to unpack why disability is best understood as a mismatch between people and systems. We explore inclusive design, responsible AI, and why average users don't exist with a focus on how continuous human oversight matters long after launch.

What You Will Learn:

• Disability is a mismatch, not a deficit.
• There is no such thing as the “average user”.
• Responsible AI means asking hard questions, continuously.
• Inclusive design must involve lived experience, not assumptions.
• Perfect inclusion is impossible, but progress is essential.
• AI can be a powerful force multiplier for accessibility.
• Everyone should define their own Responsible AI principles.

This episode explores what organizations need to get right with their data, the difference between automation and generative AI, and how agentic systems can support real workflows. We break down practical steps for readiness, governance, and using tools like MCP servers and multi-agent models effectively.

What You Will Learn:
• AI is only as good as the data behind it.
• Generative AI does not replace classic data science.
• Unstructured data is a major (and risky) part of the AI landscape.
• AI should be treated as a copilot, not an autonomous decision-maker.
• AI agents require governance and visibility.
• Not every problem needs Generative AI.
• Multi-agent systems introduce new risks.
• Education is the first layer of AI governance.