Thursday, 13 August 2026

 Nathan LAMBERT (Preston) (17:23): (1784) My adjournment matter is for the Minister for Artificial Intelligence and the Digital Economy, and the action I seek is for the minister to report back on his department’s consideration of international proposals to pace the frontier of automated AI development. I normally raise fairly specific Preston and Reservoir matters, but we did have a debate earlier this week about data centres and AI, and I want to pick up some of the points from that debate. I will not repeat the important points that were made about planning, energy use and water use, but rather I would like to pick up on this important notion that data centres are not really data centres. As we spoke about, these centres are not full of traditional file servers. They are not storing people’s data, photos or anything else. Rather, they are predominantly filled with processing units, GPUs, and they actually are the giant artificial brains that we are building to do our thinking for us when we use applications like Claude and ChatGPT. They should properly be called ‘compute centres’ or ‘inference centres’. I do not think it is deliberate, but I think calling them data centres makes them sound more innocuous and underplays just how radical this technology is.

A second thing that I think causes people to underestimate how radical this technology is is the fact that the standard models of Claude or ChatGPT that we have on our phones, whilst remarkable technology, are very neutered compared to the extraordinary capacity of the frontier models that these companies have built already, let alone the ones they are attempting to build. We have this strange situation with AI which is the reverse of what normally happens. What normally happens is that the public outside an industry are worried and the people inside an industry, be it fracking or the nuclear industry, are less worried. That is the normal course of things. With AI it is the other way round. Even the criticisms of the Greens and the Victorian Socialists are mainly focusing on things like water and energy use. They are important issues, but when you listen to some of the software engineers and researchers who have built these frontier models and used them day to day, they are seriously worried about what it will look like when we deploy at massive scale these artificial creatures that can out-think humans across a broad domain of tasks. They are worried about not only what that will do economically but what it might do in a military sense and what it might do in a very fundamental, existential way for humanity. Of course there are all sorts of positive things that can come from this. I want to be clear that I am just talking about the long-tail risks.

As someone with a little bit of work experience in this sector, the thing that I am most worried about now for my children is no longer climate change, and it is not nuclear weapons; it is where AI might go. I just encourage the other parties, the Greens and the Liberals, who I think have not yet fully engaged with this issue, to think clearly about it, as the minister and as our Labor government are doing, because I do think that this will be one of the most significant regulatory challenges that we in the Victorian government and indeed governments across the globe face in the coming decades.