Alastair Campbell

and trying to guess which one it's going to be is kind of almost impossible.

There are some things that we know we need to be very, very careful of.

AI hallucination,

is a massive problem ai's ability to do the brute force attacks on on digital security systems or you know right viruses right penetration code is really really worrying ai's inability to recognize its own errors and to lie i mean ai like there was a there was something um

where some people were testing, I can't remember which one it was, but they were testing it.

And not only did it lie, or not only did it hallucinate, when it got pulled up on it, it lied about it.

And when it got pulled up on that, it lied again.

And the thing is, of course, AI is trying to replicate humans because we've told it to replicate humans.

And what it's doing is it's replicating our faults and frailties and then amplifying them.

So we need to, you know, just like a very, very high-speed vehicle,

There is a steering wheel, but you get it wrong.

You know, you're driving at five miles an hour and you take a wrong turn, you bump into a tree.

Like, it's not nice, but it's not fatal.

If you're in a Formula One car, fundamentally the same technology, but just so much more power, it's a massive problem.

This is what AI does.

All the things that we do well, it could do very, very well and very, very quickly.

but we've got to find a way of how do we harvest the positives and mitigate the negatives?

Because what we're not going to do is we're not going to turn it off.

That's not a credible option.

Why are you a Tory?