Data democratisation works when leaders stay close to the intent, says former CIA CISO

What happens when AI makes enterprise data easier to access than it is to align?

As more organisations push to democratise data, the real risk shifts from simple exposure to fragmented decisions, weak architectural discipline, and accountability gaps that grow as access expands.

In conversation with ADAPT’s Head of Strategy Joey Meynink at the 5th Data & AI Edge, William MacMillan argues that secure scale depends on leadership presence, deliberate data movement, and shared ownership across security, technology, and data teams.

Key takeaways:

• Data democratisation creates more risk when access expands faster than leadership alignment and shared intent.
• Data movement should follow a defined decision or use case, because unnecessary centralisation adds cost, latency, and attack surface.
• Secure scale depends on shared accountability across cyber, technology, and data leaders, with architecture and governance treated as joint operating responsibilities.

Data access scales safely when leaders stay close to the intent

Opening access to data does not automatically create better decisions.

It only works when leaders stay actively involved in how that access is used, what it is meant to achieve, and how teams stay aligned as they move faster.

William’s warning is that wider access can create silent risk when teams move quickly without shared direction.

His argument is that democratisation needs consistent leadership presence, repeated communication of intent, and active validation that people understand the purpose behind the access they have been given.

Data movement should be driven by the decision being made

Many organisations still centralise data before they are clear on the decision or outcome they are trying to improve.

That slows time to value, expands attack surfaces, and creates more architectural complexity than most teams expect.

William pushes for a decision led approach instead.

He argues that organisations should test value where data already lives, using intelligence layers and guided analysis before deciding what genuinely needs to be centralised.

That creates a more selective, accountable way to move data and reduces the cost of moving everything by default.

Security and data scale together through shared operating ownership

Security problems rarely begin with a single tool.

They emerge when cyber, technology, and data teams work in sequence, hand off responsibility, or optimise for different outcomes.

William treats this as an operating model issue.

As AI increases the speed and reach of data use, those organisational seams become more visible and more dangerous.

His view is that secure scale depends on shared accountability, aligned incentives, and joint architectural ownership across CIOs, CTOs, CISOs, and data leaders.

That is what makes governance durable enough to support faster access and broader use.