Mr. Eoin O'Reilly:

I am not familiar with the Spanish authority and how it has been set up. I know from our clients that they are preparing for what is coming in the EU AI Act, so we are seeing a lot of interest in trying to understand, collate and inventory all their applications of AI. They are working through those and trying to assess the risk. In some cases, they are having to stop in their tracks in respect of things they thought they might do or would considering doing, and once they have defined that shortlist, they are then putting the processes and controls around the training and testing of those models, the quality of the data that feeds them and so on. That is very much a work in progress in anticipation of the Act going live.

On the future enforcement question, I mentioned earlier that this is a very important area, particularly in the high-risk domain, and my perspective is that it should be about more than just outlining the principles and asking people to please comply. I am happy to hear that the frontier models in the tech companies are going to have a deeper look as part of regulatory advancements in the US and into Europe, but at a local level, especially for the high-risk domains, a more formal compliance programme would be good, and I can foresee that emerging.