Salisbury

Let's get started with our first question, which is from Steve Little.

Hi.

As a 63-year-old man, I honestly don't understand if AI is realistically a threat to mankind.

As politicians, you also don't really know.

So which bot should I ask for clarity?

Right.

So I sense a certain degree of frustration in that question, Steve, and a desire to learn more.

Um, you think the politicians don't know enough about it.

Well, who think-- here thinks they do know enough about AI and wants to kick us off? That's bold, Danny.

Off you go.

Well, when you say I know enough, I mean the fact is none of us do.

Yeah.

So but I will say this, and I, and I think this is a national endeavor, and it is not an occasion for partisan wrangling.

We need to get on the same page.

The government is increasingly recognizing the seriousness of this challenge.

I am concerned that we're not moving fast enough, both with the safety agenda, and we have one of our great national assets is the AI Security Institute, which looks at these new models and sees how safe they are for human beings.

It's a, it's a world leading institute.

They haven't been given access to this latest model from Anthropic, which is a cause for concern.

But the paradoxical challenge I think we have as a country is in order to mitigate the dangers of AI, in order to ensure that it's safe for us, we actually need more of it.

We need our own computing power here in the UK.

We need to develop our own talent, which we have a lot of.

We need to ensure we have British models.

We're not so reliant on the US and increasingly on China.

So the challenge for us is we need more AI, but it needs to be safe.

This is a job for all of us.