Video transcript: Christian Arndt, Partner, Cyber Security , PwC UK and Ben Lingwood, Cloud & Data Transformation Lead, PwC UK

Playback of this video is not currently available

1:40

Transcript

Ben Lingwood: Right, so we're just back from doing our little cameo on stage talking about all things, data, cloud, and security. So let's do a quick recap first. What are your key takeaways from the session.

Christian Arndt: Key takeaways were, mainly focus on the foundations,right? You're building tech. It might be called AI, but it's built on traditional technology, so you've got to get the foundations right. Once you've got the foundations right, you really need to think about what kind of AI you're building. It's like the cloud, right?

There are lots of different phases of cloud, there's lots of different phases of AI. And they've all got different risks. So you really need to think about that. So that's my main takeaways.

Ben: Okay, nice, mine. I think the questions that we had afterwards and things that people were focusing in the session business case, it's called generative. Ai for a reason is generating something. It's not machine learning, it's not insights analytics, it's generative. So make sure that you've got that business case up front and that it's compatible.

I would say there was a big focus on the data model. To use a large language model off the shelf. Don't spend millions going to create something. Everything's going to change in 18 months anyway, right. The way things are moving.

Specifically, look at your data source and make sure that that's clean. Obviously, the liability of data as it comes through the environment. And I and then the last one operating that, you need different people, you need different processes in order to operate an AI solve a bit of a transformational change there from a perspective. Fewest key questions that came back on that.

Christian: Yeah, that's good to add on that.

You mentioned data a lot and the biggest risk with AI is you're building lots of data models where you’re storing the data.How are you securing the data? What are you putting in there?

And you can leave those data models lying around by accident, one click away from the internet. So that's my top tip.

Ben: Yeah. Good. Thank you.

Follow us