The Observation

AI adoption in the workplace creates three distinct roles and Formula 1 has spent decades refining the relationship between them. The driver executes at the limit of what is possible, increasingly alongside AI tools. The pit wall where the engineer sits makes the real-time calls, designing the strategy, decomposing complex race situations into decisions the driver and car can execute and managing the outputs. The steward enforces the regulations. In an organisation, the stewards are the governance professionals, DPOs, compliance officers, guardrail designers. Each has distinct authority, distinct accountability and a shared interest in the race running properly.

Drivers and engineers will stretch the rules. They will find shortcuts, use unapproved tools, feed data into systems that have not been assessed. That is not malice. It is human nature when a faster route appears. The steward's job is not to hand out a penalty for every incident. It is to let the race run when it produces better outcomes without real harm and to intervene decisively when it matters.

The critical shift is that the ratio of engineers to drivers is inverting. Organisations will need fewer people executing tasks and more people orchestrating AI workflows. Most AI training programmes teach people how to use the tool, which is driver skill, when they should be teaching people how to design the strategy, which is a pit wall and engineer skill.

What This Means for Data Privacy

Stewards who have never driven make poor officials. The FIA recognised this and has required a drivers' steward - an ex-racing driver - on every panel since 2010, precisely because the sport concluded that enforcing the regulations well requires knowing what it feels like to be at the limit. We often find that governance professionals who have not personally used AI tools, broken them, seen where they leak data and tested their boundaries, struggle to write policies that survive contact with reality. Spending time hands-on with AI tools, running real tasks through them and observing how they handle personal data, tends to dramatically improve the quality of DPIAs and AI policies.

It may also be worth considering whether your organisation's AI approval process reflects this dynamic. If the governance framework assumes every AI use case needs pre-deployment sign-off from a steward who has never experienced the tools, there is a risk that the process either creates bottlenecks or, more commonly, gets bypassed entirely. A governance process that deploys the safety car for every minor incident does not produce a safer race, it just neutralises everyone's progress and hands the result to luck rather than judgment.