The EU AI Act is still on track to become law. When that happens is an open-question. The European Parliament is expected to vote on the AI Act at the end of March. However, there are multiple contentious discussions that may delay progress - the definition of AI systems, which systems should be classed as High-Risk, and how to regulate for general purpose AI systems, being just a few.

Organisations looking to develop, deploy, procure and sell AI systems need to consider how to navigate the AI Act. That applies to organisations inside and outside the EU. Navigating the EU AI Act is not straightforward - the Act is nearly 150 pages long with multiple areas that are (or some may argue, could be) open to interpretation.

We have published a one-page flowchart decision-tree to help navigate the AI Act. It identifies the key decisions to be considered and references the relevant sections of the AI Act's latest draft. Of course, users should consult the latest version of the AI Act and seek legal advice on how the AI Act applies to their specific circumstances. But we hope the flowchart is a useful starting-point for structuring how they approach the AI Act and identifying areas for further discussion.

Organisations may find that they have to navigate the EU AI Act as well as regulations in other jurisdictions. How they comply with those regulations (and other relevant laws) may look very different to their approach to complying with the EU AI Act. For example, you can see the different approaches being taken by looking at: a) our one-page visual on anticipated AI regulations in the UK, EU and US, see our horizon scanning (click here); and b) our glossary of existing and anticipated AI definitions in UK and EU regulation, legislation and policy click here.

Engaging with anticipated regulation early is therefore essential; organisations may have to change their business model, product offering, technical processes and, in any event, approach to governance and compliance.

Click here to view the flowchart.