澳专家呼吁AI监管：美政客忽视风险，中小国可联合定标

Around the debate surrounding AI safety and safeguards. President Trump dismissing industry calls for a slowdown as part of a conspiracy theory. China also calling fresh warnings about the tech fearmongering. It comes as the sector considers establishing common safety standards. Microsoft has now released an AI manifesto that it says puts people first. After the CEOs at anthropic, OpenAI and SpaceX, I endorsed voluntarily slowing AI advancement.

Let's get some more with Toby. Walsh is the associate professor of artificial intelligence at the University of New South Wales here in Sydney. Really great to have you with us. And it comes on a day that we heard from, uh, Australian leader, the signals directorate director general, saying that we need an early warning system for I hear. I guess my question to you is multifold. Do you think safeguards need to be in place?

Is it a little bit late in terms of the progress that's already been forged ahead with, and who should be putting in these safeguards? Given the difficulty I think regulators have of understanding and keeping up with the tech. Come on. It's very probable that is going to kill us all. But I starting to have the capability to do some serious, ah, particular things like the cyber risk. That was actually a very troubling report from anthropic last week, showing that Iranians have tried to do bio weapon research using the closed chat bots, and Russian hackers had tried to build some autonomous swarms of kamikaze drones.

Those are things that we started to be worried about. And as the chief advisor says, um, it really is something that we need some warnings about. At the same time, of course, the US has said it's really improving lifting their game as well. For Australia, what steps should be taken? Because this is happening at the same time, when you have really both sides of government trying to work out how broadly the economy can take advantage of the AI opportunity, whether it be the build out of data centres, whether it would be taking up the opportunities that we've heard from some of the AI giants of training models in Australia.

Yeah, I mean, it is actually keeping our economy afloat at the moment. Latest ABS copy figures, uh, said that if it weren't for the spend on building AI data centres and the like, we would have gone into recession. So the government's obviously right. Be very keen to ensure that we do get the benefits we do that kept the growth. But at the same time, uh, we also need to think about some of the risks, especially the longer term risks.

Uh, and those are things not just national issues, though. I mean, this is an international race. We've got to work out how we coordinate, um, how the US and China actually manage these risks as well. And it's a bit troubling that Trump is dismissing the risks completely. I like that, you know, as you say, the good news is I very much doubt we're all going to die because of rogue. I mean, that's a pretty low baseline, right?

We could see some pretty seismic impact when it comes to the economy. When it comes to financial assistance, when it comes to the social fabric. Do you worry about these issues? And I guess a concern is who's in the right place to be deciding what the guardrails and limitations should be at any given time? Yeah, very much so. I'm actually much more worried about some of the longer term or insidious risks that Ross is going to have on people's employment.

Is it going to create an underclass of people who are put out of work, and what's the impact it's going to have on our political discourse? We've already seen what social media did to that. It polarized us and turned politics into a. But first letter I want is going to have to our relationships where increasingly people are starting to have relationships, synthetic relationships with parts. 200,000 teenagers in Australia have had an I boyfriend or girlfriend.

Uh, I worried what the impact that's going to have on us as humans in the long term. So there's a there's a lot to worry about. And so not that our government should be doing in terms of protecting its citizens, as well as seizing the economic possibilities. Do you think it's particularly worrying that we're not hearing from the top leaders of these AI companies? It's not just whistleblowers, right? It's not just grassroots activists that are now concerned.

When you hear from the likes of the heads of anthropic, of OpenAI, from Elon Musk. Does that really suggest that there is an existential threat here, that they would want to slow the pace of development? That you suggest we should be thinking about slowing the pace of development? Uh, Elon Musk, Sam Terrio, Modi the heads of these companies hate each other. They agree on almost nothing. So the fact that they can agree on this is actually quite telling.

Um, and so I think we should I mean, it's it's probably been a little late in the day for them still to be marking their own homework. They've actually asked, uh, suggested that we should have independent government oversight, audits, auditing what's going on in labs. We do that with drug companies. We do that with aviation companies. Uh, we do that in any other business. But where there are significant risks to human life.

I starting to play a business in which the significant risks, possibly to human life. Uh, I think there needs to be better oversight. We can't rely on that. Good. Well. I love the mocking your own homework comparison, Toby. You know, uh, do you think the expertise is there in the public sector in terms of independent regulators? Uh, that's a great question, because you're right. Um, we're talking about pretty capable, competent people who can command very large salaries, uh, in the private sector.

So it's going to be hard for government to do that, though they've already got on the I mean, most UK government, the United States government here in Australia, we've set up these safety centers. Um, there's there's significant expertise. They maybe need some more resources. Um, and they probably need some more law to, to add to what they do, but the governments have started to move down the right road. It's quite interesting for all of the feuding.

Uh, and of course, we're expecting, ah, discussions to take place when Presidency and Trump meet a little bit later. Both seem to be on the same track when it comes to dismissing these concerns as fear mongering. President Trump in his, uh, social media, um, statement saying it's a hoax. He thinks it's a conspiracy, right? This idea that it needs to be slowed down if the two biggest economies and two biggest tech ecosystems in the world are going down that track, what should middle countries like Australia be aiming for?

Do you think there's value in a coalition of smaller countries that might want to reach some degree of of, you know, conservative guardrails for artificial intelligence? Uh, 100%. I'm backing. Uh, people like Mark Carney. I think he's completely correct. Middle class. Like, um, Canada, Australia and those powers like Europe, uh, could really set the agenda here. It's really important that we work together with those mental powers with Europe.

Um, we're actually in Europe, plus Canada, Japan, Australia, um, is actually a larger economic zone, uh, than the United States or China on their own. So. So, uh, we've got a significant role to play, and we should realize that together we can actually really set the norms and standards for the rest of the world to be really great to have you with us. Toby Walsh, professor of artificial intelligence at You NSW, Sydney.

We'll have more on Australia head every Tuesday at 10:40 a.m. if you're watching in Sydney. 8:40 a.m.. If you're catching us in Hong Kong, you can also tune in to the Bloomberg Australia podcast, delving into the bigger stories shaping the country's role in global business. You can find it on Apple, Spotify or Bloomberg.com. More ahead here. On the Asia trade.