So could you give us a very basic introduction to the EU's AI Act? And as part of that, can you help me understand why this Act seems to have got so much attention?

Over to you, Sue-Anne.

Thank you, Therese, for asking this question, because I think it sets the scene in terms of looking at AI regulation in general, but also bringing forth the European approach.

So the EU AI Act is a milestone in a sense, because it's the first in the world to comprehensively regulate AI.

It treats AI as a distinct object of regulation.

Using this thing called a risk-based approach and which treats AI as a product from a product safety legislation perspective.

So the risk-based approach means that the type of obligations that are set out for those who design and deploy AI depends very much on a risk classification.

Are human rights protected by the EU's AI Act?