Episode

Handling personal data is a privilege, not an entitlement.

Sona Mehta, data protection lawyer and senior lecturer at City St George’s, University of London, on the gap between policy and practice, treating personal data as a privilege, and the checks to run before you rely on a third-party AI model.

Published

16 November 2026

Duration

33 minutes

Guest

About the speaker

Sona Mehta is a commercial lawyer who has worked in data protection since the dot-com era. She is a senior lecturer at City St George’s, University of London, a consultant solicitor and a non-executive director, and teaches and speaks on how privacy really works inside organisations.

Episode guest

In this episode

On paper, the dots are joined. In practice, they’ve moved.


Sona’s answer to the gap is simple. People assume what’s written down reflects reality: the dots are all joined and everything lines up. In practice, it rarely does. Systems change, new tools arrive, people come and go, and now AI has come in on top. The paper version stays where it was.


She sees the same pattern as a lawyer, in what the law says versus what actually happens. Organisations write things down first, then get on with the day job, and what’s on paper quietly becomes history.


“What’s on paper is historic. People write it down, get on with their day jobs, and practice moves on without it.” - Sona Mehta


What has changed in twenty years is awareness. Data protection used to be a tick-box exercise: a privacy policy had to exist, but nobody read it. Today, people know their rights and recognise a data protection issue when they see one. The gap now isn’t knowledge. It’s keeping the paper in step with how the organisation actually works.


Handling personal data is a privilege, not an entitlement.


For Sona, the UK GDPR baseline comes down to one thing: the focus is on the individual. That means collecting only what you need, telling people what you’re doing with their data, not keeping it longer than necessary, and keeping it secure. From the European point of view, being allowed to handle someone’s data is a privilege a business has to earn.


She gives an example from policing. Picture every lever-arch file on the shelf as a victim. If one goes missing, the damage to that person could be serious. Once you see data as people rather than numbers on a spreadsheet, you treat it differently.


The same applies to where data lives. Ask how the business would cope if it couldn’t reach its data, then choose where to store it with that in mind. But keeping data on site isn’t a fail-safe either. Wherever it sits, it needs the same controls.


“It’s a privilege, not an entitlement. If you start thinking of data as people, not numbers on a spreadsheet, you treat it very differently.” - Sona Mehta


Know what the model learned from before you trust it.


Before relying on a third-party AI model, Sona’s first question is what it was trained on. Was the data scraped from the web, given freely, or bought from brokers? Did the people in it agree? Is it skewed towards one group, such as graduates of the same few universities? If it is, those biases won’t disappear in your use. They’ll be amplified in your results.


Getting comfortable at the start isn’t the end of it. You have to keep checking the model over time. You also have to ask what happens to your own data once it goes in: does it become training data for the next customer?


Her one change for leaders follows from this. Bring your data protection lead into the strategy from the start, not the night before a contract is signed. Brought in late, they see only risks and become “the no people”. Brought in early, they help the business get there faster.


“Involve your data protection lead from the outset, and they stop being the no people. You get to the result quicker.” - Sona Mehta

More

More from the series

Worth your time – security, privacy, and resilience leaders across Europe.

When Controls Meet RealityOperational Resilience, DORA, Risk Management, Financial Service
FriSept 2026

When Controls Meet Reality

A framework that reads well for auditors can still fail when pressure hits. In this episode of Montro In Practice, Ankur Arora sits down with Emmanuel Ranchin, VP of Information Protection Governance at State Street, on the gap between documented controls and lived ones: why compliant isn't the same as resilient, how big institutions break where silos meet, and what DORA really asks of the board.

Watch now
View all

Don't miss the next conversation

Get new Expert Series episodes and governance insights in your inbox.

Browse episodes