Artificial Intelligence

David Shoebridge

I seek leave to move a motion relating to the regulation of AI and threats of extinction, as circulated.

Leave not granted.

Pursuant to contingent notice standing in the name of Senator Waters, I move:

That so much of the standing orders be suspended as would prevent me moving a motion to provide for the consideration of a matter, namely a motion to give precedence to a motion relating to the regulation of AI.

I can already almost hear the interjections from the chair here: 'You need to talk to urgency. You need to say why this motion is more urgent than the other business that the Albanese Labor government has before us.' So let's look at the great agenda that the Albanese Labor government has today: the Regulatory Reform Omnibus Bill—that's what they say is urgent to get to—or the Major Sporting Events (Indicia and Images) Protection Amendment Bill 2026. That's what's up for us this morning—not talking about how this parliament, this government and this country should be moving to address an extinction-level threat to us as a species coming from AI.

I'll tell you what the Greens think we should be talking about. It's not the Regulatory Reform Omnibus Bill, but the fact that major players from the AI industry across the world are saying we need to slow down the development of frontier models because it may make us extinct. This isn't some kind of far-right conspiracy theory on the internet. Jacob Coxon quit Anthropic, one of the more ethical players in the space—if you can call them ethical. He accused Anthropic and AI of 'racing straight to self-improving superintelligence and gambling with our lives'. Then he said:

The people building AI earnestly believe that it could kill us all by the end of the decade.

That's what the Greens want this parliament and the Senate to be dealing with, not what Labor's got on the books, which is the Regulatory Reform Omnibus Bill.

It's not just Jacob. We could look at what Evan Hubinger, who's still at Anthropic, said. He came out just this week about the unlimited development of frontier AI models—which, to be clear is Donald Trump's plan. They're the models that will run on the AI data centres that Anthony Albanese, the Prime Minister, wants to put across our country like mushrooms. Evan Hubinger said that he personally believes there is a chance, in the order of 10 per cent, that AI could cause human extinction within the next decade. The Greens are listening to this. A 10 per cent chance of human extinction is a reason for urgent action. That is why we've brought this motion before the parliament. That is why we're saying act on this now. Our motion notes that major global AI corporations, multiple whistleblowers from big tech, are telling us that unregulated AI is a near-term extinction risk and that urgent guardrails are needed.

We're calling on the Albanese government to legislate mandatory safety testing and transparency obligations for frontier AI systems, rather than relying on industry self-regulation. We're calling on the government to urgently reach out to middle powers—starting with Canada, Singapore, Japan and South Korea—and push for a global treaty on AI regulation.

We don't pretend that Australia can do this by itself, but I'll tell you what we think we should do. We could lead the global debate here. We could bring forward legislation to put mandatory guardrails around AI. We could set out the rules for an ethical AI industry that constrains the frontier models, puts them back in the box and makes sure the checks are done before they're rolled out and sent into the wild. We need to urgently work with other middle powers to push back against the likes of Donald Trump. Other countries are calling out to us to join them in global leadership in this space. Canada is calling out, and, in our region, South Korea, Japan and Singapore are calling out. We should join with these other middle powers to do what is essential now for the future of our species and for the future of our jobs—to actually put in some global guardrails.

I'll deal with a couple of other comments. Dario Amodei, the CEO of Anthropic, has said that there is a warning of an AI-agent swarm that would become capable of taking over the entire internet within a year. That's what Anthropic is telling us—an AI-agent swarm taking control of the entire internet within a year. Even Sam Altman and Elon Musk—the crazy far-right end of the industry—are saying it.

I can't believe anyone would vote against this motion. I can't believe Labor would say: 'No, we don't want to deal with an extinction risk. We want to deal with the Regulatory Reform Omnibus Bill 2026.' But— (Time expired)

Tim Ayres

Let me make a few observations about this particular proposition and then about the government's approach to this set of questions. The first is that we heard a lot from the Greens political party last week about the sacrosanct nature of Thursday afternoon debates. You don't do the Senate any favours with 10 minutes notice of a—

I don't want to contradict the smartest senator in the room.

Slade Brockman

Order, Senator Shoebridge!

Tim Ayres

I don't want to contradict galaxy-brain over there.

David Shoebridge

Through the chair, mate! Through the chair, buddy!

Slade Brockman

Minister, it probably would help if you directed your comments through me.

Tim Ayres

I should, Deputy Chair, and I will.

Slade Brockman

Thank you.

Tim Ayres

If there were any political seriousness at all from the Greens political party and not a desperate attempt to cover over their catastrophic last couple of days, then there would not be 10 minutes notice of something that they claim is a species risk. These two propositions cannot sit alongside each other. This is just about political convenience. It is not about the substance of these questions; their approach on these issues never has been.

Slade Brockman

Senator Shoebridge, on a point of order?

David Shoebridge

The minister is clearly not speaking to urgency. That's the debate.

Sue Lines

Minister, I will bring you back to the debate before the chair, which is the question of urgency.

Tim Ayres

I would say, on the point of order, that, obviously, credibility goes to urgency. What I'm indicating is it is not a credible position—

Slade Brockman

Minister, I'll rule on the point of order. I will remind you that we are debating the suspension motion. It's important that we do relate our comments to the suspension motion; however, you have the call.

Tim Ayres

It is in no way a proper use of the Senate's time now to deal with this question. It is a crude and base political stunt. It is.

Slade Brockman

Order! Senator Shoebridge, you were heard in silence!

Tim Ayres

Senator Shoebridge has just given the game away. What he was talking about was the clip. What is this about? It is about the video clip that poor old Senator Shoebridge is desperately hoping that the cover up of the woes of the Greens political party is a social media clip. Senator Shoebridge and others—and I largely agree with them—carry on about the corrosive impact of social media in our political discourse. I agree, but you're part of the problem. This Senate institution—

Slade Brockman

We are getting a long way from the motion.

Tim Ayres

The motion goes to urgency.

Slade Brockman

Yes, thank you for bringing us back there.

Tim Ayres

Senator Shoebridge says this is about video clips. He has—

Slade Brockman

Senator Waters, a point of order?

Larissa Waters

A point of order on reflecting adversely on a senator. Reel it in, and come to the topic at hand. You're entirely unprepared on this issue, and it shows.

Slade Brockman

Senator Waters, resume your seat. Minister, you have the call. Direct your comments through the chair.

Tim Ayres

I agree with some of Senator Waters's observation there too. Senator Shoebridge gave 80 per cent of the game away, and Senator Waters gave the last 20 per cent of the game away. It's about clips, and it's what was called, in that great COVID era show—

Slade Brockman

Senator Shoebridge, on a point of order?

David Shoebridge

This minister could, at some point, address the substance of the issue. Either talk about the extinction risk, which he's refusing to do, or talk about urgency and why the Regulatory Reform Omnibus Bill is more important.

Slade Brockman

Senator Shoebridge, resume your seat. Minister, you have the call.

Tim Ayres

In that great COVID era show Kim's _C__onvenience_, it's what the owner of the convenience shop called the sneak attack. From Senator Waters it's, 'We'll bowl up five minutes before with a brand new proposition.' It is all about the most facile politics. What the Albanese government is—

Slade Brockman

Minister, resume your seat. Senator Allman-Payne?

Penny Allman-Payne

On a point of order, the senator is making imputations about the motivations of senators. He needs to speak to the urgency, or his argument is not urgent.

Slade Brockman

Senator Shoebridge, don't interject when I'm making a ruling. Senator Allman-Payne, there was no imputation. I will bring the minister back to the question at hand.

Tim Ayres

The approach from the Albanese government is very straightforward. It is to make sure, not just for economic reasons but for strategic reasons and for the safety and long-term agency and resilience of the Australian people, that we capture the right artificial intelligence capacity in Australia for our agency and for our capacity for Australia as a nation, not just now but into the future; that we make sure that we spread the benefits of this technology right through our community; and that we are completely focused on safety.

The Artificial Intelligence Safety Institute is doing its work. It's engaging with many of the countries that Senator Shoebridge's resolution touches on to get under the bonnet of these frontier models and to engage with our intelligence communities and others to make sure that artificial intelligence in Australia is in the national interest and that our AI standards are— (Time expired)

Mehreen Faruqi

This matter is urgent. An Anthropic researcher has just quit, warning that AI could bring catastrophic harm to humanity. A tiny club of depraved tech oligarchs, billionaires and corporations accountable to absolutely no-one are racing to build technology they themselves warn could destroy millions of jobs, upend democracy and even threaten humanity itself. This is insanity, and we are letting it happen.

These threats are not hypothetical. These threats and risks are urgent. AI is already being weaponised for warfare, surveillance, deepfakes and sexual abuse, while its sprawling data centres swallow staggering amounts of water, energy and land. This machine is already chewing through the planet and giving powerful people terrifying new tools to exploit and control us. Nobody voted for this. Nobody wants billionaires gambling with humanity's future. Nobody wants data centres sucking up water and energy while the planet burns and floods. Nobody wants every click, every conversation, every photograph and every detail of their life harvested for profit. Nobody wants a handful of self-interested men deciding what our future looks like, yet governments are rolling out the red carpet.

Labor will regulate ordinary people and punish protesters, but when they're confronted with corporations and billionaires wielding the most unimaginable wealth and power, with technology capable of unimaginable destruction, suddenly they turn to water, suddenly their courage disappears. We should be afraid and we should be angry. But, most of all, as Naomi Klein says, we should build a wealth and tech disarmament movement. That's what the Greens are doing, because we cannot let oligarchs own our future. Humanity is not a billionaire's experiment.

Slade Brockman

I will remind senators we do need to be relevant to the suspension motion.

Long debate text truncated.