Mr. Dualta ? Broin:

I thank the Deputy. I have linked to our adversarial threat report in my opening statement and it covers generative AI and goes into the types of trends we are seeing in that. The technology has improved and our systems are scaling and are scalable further to deal with generative AI as a threat. A lot of the co-ordinated inauthentic behaviour type, which is where we might see this generative AI being used, is based on behaviour rather than the content itself. To date, the content itself has not necessarily proven to be the thing that helps us remove networks of co-ordinated inauthentic behaviour. The second thing is that we have found that while they have capabilities on content, the capability to doctor images goes back, though I appreciate it has taken a leap in the last while, and the real challenge for the bad actors trying to basically upset the democratic processes in various countries - and some of them succeed - is around the behavioural elements. That means getting accounts to be able to act as the conduit for that content. The content to date has not necessarily been an issue.