Senate Foreign Affairs, Defence and Trade Legislation Committee, , 8:44:14 to 8:47:54
Discussion on the ADF's responsible use policy for artificial intelligence and autonomous weapons systems, particularly regarding lethal payloads.
Topics: Artificial intelligence, National security
Speakers: Hugh Jeffrey, David Shoebridge, Meghan Quinn, Speaker 32, Unknown Speaker, Interjection, Raff Ciccone
>> We know there are plans.
>> Last question before I move the call around. S
>> we, as I understand it, there are plans to attach um drones and mines to the ghost shark.
>> Different payloads. Yes.
>> Different payloads. Some of those will be lethal is my understanding. Mines, for example, would be lethal. Is that correct?
>> Potentially. Yes. The intention is for it to have payloads that do different effects. Yes.
>> All right. Um um given that it's a partly AI platform, given it's operating remotely, um and given that its payloads are intended in the future to be lethal, does the ADF have any formal policy around how to manage AI and autonomous weapons? And I I note that as we're talking about Andrew, Palmer has made it pretty clear that he is very supportive of fully auto automated weapons um being delivered by Andrew. So um does the ADF have a policy? If so, where do I find it about managing AI and autonomous weapons?
>> So we do have a responsible use policy related to AI across the enterprise including in our capability platforms. So
>> yeah, but my specific question is about autonomous weapons, AI and autonomous weapons.
>> So So the responsible use of AI applies to everything we do, including autonomous uh develop autonomous
>> Uh Senator Chris Crozier, I'm the chief artificial intelligence and information officer. Yes, there is a responsible use policy that's been uh published uh earlier in the year which does provide a comprehensive overview of defense's position.
>> And what does it say about AI delivery? AI killing people.
>> Pardon me. Sorry.
>> What does it say about AI killing people? AI delivering lethal force.
an international uh the responsible use um uh uh for AI is governed by the policy. The policy is published online. Uh this is a uh uh this is required by government for the adoption of artificial intelligence into military system. Senator, as you know, uh, artificial intelligence will transform warfare and is already doing so. Um, and the big risk for Australia is that we face ongoing and increasing increasing asymmetric disadvantage uh by potential adversaries who are increasingly integrating AI into their kill chains. We need to keep pace with that. However, we need to do it in a way that is consistent with Australia's international, domestic, constitutional, legal obligations. Uh, and so the purpose of the policy is to ensure that Australia can integrate artificial intelligence into military capability according to the principles set out in the policy. Those principles, Senator, u the most important among them must be uh human in the loop to ensure that the allocation of force, the targeting of assets is always an act authorized by human in the in the targeting system. uh and uh the how we integrate into the force is uh the responsibility of course of our um