Navigating the AI Maze: Copilots, Data Security, and the Risk of Getting Dumber

Introduction

In this episode, we talk with Marius Sandbu about Copilots, licensing, data silos, and the question of whether technology is making us smarter — or more dependent.

Meet the Guest

Marius Sandbu specializes in Microsoft cloud, security, endpoint management and modern workplace technology. He's active as a speaker, consultant and community member.

Setting the Stage

Microsoft uses the name "Copilot" for many different products. That raises questions about functionality, data, licensing and how the various Copilots work together.

Episode Highlights

• Microsoft now has dozens of Copilots, and the list keeps growing.
• AI can strengthen human skills, but it can also erode them.

Deep Dive

When every Copilot operates in its own silo, new complexity emerges. A Copilot for Intune and a Copilot for Security may look at overlapping data, but they don't automatically work together. We also need to think about human skills: the more technology does for us, the less we practice certain skills ourselves.

Real-Life Stories & Examples

• Copilot for Intune and Copilot for Security look at the same environment from different perspectives.
• Parallel parking without cameras or sensors illustrates how skills can fade.

Key Takeaways

• More Copilots doesn't automatically mean less complexity.
• Product names, licensing and data sources need to be clear.
• Copilot silos can create new information barriers.
• Retain foundational knowledge for situations where technology fails.
• Use AI deliberately.

Closing Thoughts

The Copilot maze becomes manageable once organizations properly map out applications, data and responsibilities. AI should help us achieve more — not make us forget how to think for ourselves.