What really changes in your organisation with the AI Act

The entry into application of the European Artificial Intelligence Regulation does not by itself answer the question most organisations are asking: what do I have to change, in what order, and what evidence shows that I have done it.

The AI Act classifies systems by risk level, but classification is only the starting point. The same model can be limited-risk in one context and high-risk in another depending on who it affects and which decisions it conditions.

Three fronts to open at the same time

The first is the inventory. Many organisations discover, when they do it, that they use more AI systems than they thought, many of them built into third-party tools.

The second is governance: who decides, who oversees and who is accountable when a system fails. Without assigned owners, documentation becomes a paper exercise.

The third is training. A regulatory framework only works if the people who work with the system understand what they are looking at.

Sample article for blog layout.