Could AI Show the Wrong People the Wrong Data?

AI can be a huge productivity boost, but only if it can’t see (or share) information it shouldn’t. 

One of the most common concerns we hear is simple and very real:
“If we turn on AI, could it surface things like salaries, HR data, or confidential documents to the wrong people?” 

In this session, we’ll break down, how Copilot decides what it can see, why oversharing happens so often in organisations, and how small gaps in permissions, labelling, or retention can quickly turn into real risks once AI is introduced. 

We’ll walk through practical examples (including sensitive HR and finance data), explain where organisations usually get caught out, and show how Microsoft Purview helps you: 

• Understand who can see what today 

• Reduce oversharing without locking everything down 

• Apply the right protections so Copilot only surfaces appropriate information 

• Prove control for audit and compliance teams 

This is not a technical deep dive. It’s a practical, business‑focused session designed to help you enable AI with confidence, avoid awkward (or serious) data exposure issues, and make sure security supports adoption rather than blocking it. 

Perfect for anyone responsible for approving, governing, or supporting AI, whether you’re security‑focused, data‑focused, or just want reassurance before switching it on.