Andrew Williams

Hi, Ed, how are you going?

Yeah, really good.

Good.

Now, you've said previously that around 80% of AI harms could be addressed using the laws that we've already got.

What does that look like in practice and where are the gaps at the moment?

So with AI, the systems can be quite opaque and they can be difficult to understand.

It's not always clear who's responsible when something does go wrong.

How should we be thinking about responsibility in that context?

Right.

One of the issues with AI that sort of popped its head up a lot in recent years has been worker surveillance, AI being used to actually monitor employees.

You've released a report on this.

What's actually going on for those that might be worried while they're working in their office that they're being looked at by AI this whole time?

So just stepping back to AI more broadly, what are the practical steps that governments might be able to take to try and reduce harm now without bringing in a whole bunch of new frameworks or regulations?

Yeah.

Ed, it's an absolutely fascinating subject.

Thank you very much for your time.

Thank you for having me.

For more, you can head to safeai.org.au.

This interview was brought to you by Mindaroo.