Could artificial intelligence be key to improving Australia’s living standards?A Productivity Commission report suggests the roll out of the technology could inject more than a hundred billion dollars into the economy over the next decade. But there’s a catch. If AI can take over from workers on some tasks, are swathes of humans set to lose their jobs? Today, Toby Walsh, the chief scientist at the AI Institute at the University of New South Wales, on whether it's time for the government to step in. First published 1 September, 2025Featured: Toby Walsh, Chief Scientist at UNSW.ai, the AI Institute of University of New South Wales SydneySubscribe to ABC News Daily on the ABC listen app.Program:More from ABC News DailyAustralia, Technology, AI, Copyright, Copyright Infringement, Business and Industry Regulation, Regulatory AuthoritiesTranscriptSam Hawley: As we've discussed before, Australia has a productivity problem and the government wants to boost it. Well, artificial intelligence could be the key. According to one report, it could inject more than $100 billion into the economy over the next decade. But there's a catch. If AI is more productive than you, you might lose your job, and some people already are. Today, Toby Walsh, the chief scientist at the AI Institute at the Uni of New South Wales, is going to give us some insights on whether it's time for the government to step in. Toby, there are all these tech bros sitting around in Silicon Valley and, you know, they're doing some pretty amazing things with AI. The extent of it is pretty extraordinary, isn't it? Just tell me, where are we up to when it comes to AI? How much can it actually do at this point?Toby Walsh: It can do an increasing amount of people's jobs. Equally, it's worth pointing out there's still some significant limitations. It doesn't have our emotional intelligence, our social intelligence, our common sense. But there's lots of routine stuff that we do day to day, whether that be graphic design or answering customer calls, AI can increasingly do a decent job.Sam Hawley: Well, back in 2023, one of these tech executives, the head of OpenAI, Sam Altman, he actually testified in Congress, warning that things could go, in his words, quite wrong because AI could see the elimination, in part, of millions of jobs. That was his view back then.Sam Altman, OpenAI CEO: Look, we have tried to be very clear about the magnitude of the risks here. My worst fears are that we cause significant, we, the field, the technology, the industry, cause significant harm to the world. I think if this technology goes wrong, it can go quite wrong.Sam Hawley: People were pretty concerned, and that was only a few years ago, right?Toby Walsh: Yes. I mean, it's huge uncertainty what the net impact is. I mean, technologies have always taken jobs away. Technologies have always created other jobs at the same time. In the past, technology has always created more jobs than it took away. There's no guarantee, of course, that history will continue to repeat itself because we are taking away some pretty unique capabilities, and machines are able to do those. So it may not be the case that we have as many jobs. We may end up working a shorter week. We may have fewer people in employment. It's hard to be sure. I think what is certain, though, is that it's going to be disruptive. And even if there is many jobs, or more jobs created than destroyed, people are going to have to find new jobs, and they're perhaps going to have to have new skills for those new jobs.Sam Hawley: Yes, things are moving very quickly. Sam Altman, too, he's changed his tune since 2023, and since Trump came back to the White House, he's really saying that, you know, just let AI rip, basically. He doesn't think that you need laws that would actually hinder the advancement of AI.Sam Altman, OpenAI CEO : I think that would be disastrous. I am nervous about standards being set too early. We need the space to innovate and to move quickly.Toby Walsh: Well, mostly that letting it rip is because they want to compete against China and don't want to be held back by any laws. It doesn't mean necessarily that it might not be for the benefit of most of us if there were a few rules in place.Sam Hawley: Just tell me, though, before we move on, Toby, what sort of jobs are most at risk from AI at this point?Toby Walsh: Well, certainly any job that you're doing something very repetitive and dull, that's likely to be the things the machines can do. And interestingly, I think most people suspected a few years back that it was going to be, you know, jobs in factories, that robots would be coming in and doing those. It turns out, actually, there's quite a lot of things that people do in offices that are actually perhaps easier for machines to do than they are for factories. Robots are quite expensive, and lots of moving things around in offices, moving the mouse and the keyboard around in the office or answering the phone, those are things that machines are doing. So it's actually some of the jobs in the middle, whether that be, you know, a customer service, a graphic designer, or even entry-level computer programmer. That's, I think, surprised lots of people, that that turns out to be something that computers now can do quite well, and that we thought, you know, they were going to be the jobs that would be safe in the future, and they turn out to be not the case. Or even, I mean, this is one that got surprised me, was fashion model. I thought, well, there's no way we're going to have robots walking down the aisle, but it turns out that there's a lot of stuff that fashion models do that's digital. It's, you know, making catalogues and seeing, you know, different size models, different colour models with the clothes on, and you can get a generative AI to make those photographs very realistically.Sam Hawley: Gosh, and make them as attractive as you want, I assume, which might not be such a good thing anyway. Toby, let's consider now the idea of laws, or further regulation, to protect jobs from being taken by computers, if you like. The unions, of course, are really worried about this. The ACTU wants something done to protect workers.Sally McManus, ACTU Secretary: Why do you need more powers when it comes to AI? What we are saying is that you should consult with your workers and reach agreement on the things that they're going to really be worried about, like their jobs and are they going to be retrained, and what about their privacy and their data, when you make the decision to introduce AI.Sam Hawley: But do you think that's possible? Can we protect workers against AI at this point through regulation and laws?Toby Walsh: I mean, it is possible to work with unions, whether we need actual regulation, formal regulation to do that, or whether it's about the companies maintaining good relations with their employees, realising that the most valuable thing that most companies have are the people that work for it and the skills that they have and the knowledge that they bring to their day-to-day job. We've seen it in the past. I mean, for example, Rio Tinto, when they automated some of their mines, they took out all the human-driven trucks and replaced them with computer-driven trucks, and there were 300 odd drivers who were impacted by that. That was great news for their bottom line. They didn't have to pay those truck drivers. Great news for safety. The mines were much safer, with the computers making fewer errors than the humans were. But they did work. The management of Rio Tinto did work with the unions and promised that every one of those drivers would be found alternative employment somewhere else in the business. That's certainly possible for a big business like Rio Tinto, where they employ thousands of people, that they can find alternative employment. But I think that was a good example of where it's done right. How we ensure that happens elsewhere, that's an interesting, challenging question.Sam Hawley: Oh, yes, Toby, an interesting case in point. The Commonwealth Bank, after pressure from the union, it backflipped on its decision to axe about 45 jobs, which were actually going to be replaced by AI, and it even apologised to workers because of it.News report: CBA now says they didn't adequately consider all business decisions and an error meant these jobs actually weren't redundant. The union is describing this backflip as a major win for workers, but it says that many jobs here in Australia are still under threat from AI.Sam Hawley: Well, that's why, of course, the unions want a national AI act, just looking specifically at AI. Other nations have done that, right? The European Union has done that.Toby Walsh: They have, although most of, you know, if you look at, for example, Europe, the AI Act that exists there, there isn't very much in that that is about protecting people's employment. It's more about protecting the public, thinking about perhaps the new harms that these technologies would bring up in terms of subliminal advertising and manipulation of what people do. The rights we have as consumers, very little really in the AI Act about protecting people's rights and not preventing them from being replaced by machines.Sam Hawley: A big concern, of course, is the creative arts sector because artists are really vulnerable here, aren't they? Because AI can easily be trained off their work. Surely they do deserve some better protection.Toby Walsh: Well, they do, and we already protect artists' work with things like copyright, and there's certainly a very vigorous ongoing debate about whether what the tech companies have done by training upon those artists' work was actually fair use or not. There are multiple lawsuits in the US and elsewhere arguing that they should have sought the consent at the very least and perhaps offered compensation to the artist whose work was used in the training. And indeed, many of those tech companies are now starting to do licensing deals with some of the big media conglomerations, with Nine, for example, the New York Times, to actually license the data that they've been trained on to ensure that it was fair use. And I think there's a cultural argument here. If we allow the income of these artists to be taken away, because now all of us are using generative AI to do our graphic design and we're not paying graphic designers, for example, whether that's sustainable, what will happen to the culture that we value and the artists who depend upon having some income?Sam Hawley: All right, well, Toby, of course, on the flip side of this argument, is concerned that regulation, that laws could actually be quite damaging. The Australian Productivity Commission, it's released a report. It's really worried that too many restrictions on AI could really dent our economic growth.Danielle Wood, Productvity Commission Chair: While managing risk is important, we do not think that a new and overarching framework for AI is the way to go. That's because the risks posed by AI are mainly existing risks.Sam Hawley: It's predicting that over the next decade, AI could inject $116 billion into our economy.Toby Walsh: Yes, I mean, we see some pretty big numbers. I mean, they're often around, you know, adding 15% to our GDP. And I think that's a fantastic opportunity. I mean, there's few other opportunities that size coming along that can help us grow our productivity, can contribute to improving our tax base and the GDP of the country than artificial intelligence. So we should do everything to ensure that we do get that benefit. But equally, we want to make sure that, you know, that benefit doesn't just flow to these tech companies in California and elsewhere. We want to make sure that we have a healthy, vibrant AI industry here, that we're using the technology and getting value out of it ourselves. It's not all being exported to overseas.Sam Hawley: Yeah, well, the author of that report, Stephen King, he likened AI to the way steam engines helped begin the industrial revolution.Stephen King, Australian Productivity Commission: Whether it's research, whether it's medicine, whether it's communications, whether it's farming, we've seen AI being used in farming to help fruit sorting, whether it's mining. Any area of the economy is going to be improved by what is a general purpose technology, AI.Sam Hawley: So they would just need it and you don't want to stand in the way of it.Toby Walsh: You don't, but we also, we did actually introduce quite a bit of regulation around the industrial revolution to ensure that the benefits were distributed and shared amongst all of us and not just by the, as they became known, the robber barons of the time. We did introduce unions to fight for the rights of the workers. We did introduce labour laws to ensure that we didn't send children down the pit. We did introduce a welfare state to support people when they were out of work. We introduced universal education so people were educated for those jobs. We introduced universal pensions so people could retire at the end of their working life. So we did a lot of changes, a lot of regulation, actually, to ensure that the benefits were spread around and it wasn't just the owners of the means of production that got those benefits.Sam Hawley: The Treasurer, Jim Chalmers, he's also urging against a heavy-handed approach to regulation.Jim Chalmers, Treasurer: My view is that we can find a middle path between people who say, let AI rip, and other people who pretend that we can just turn back the clock. I think there is a middle path here and when it comes to regulation, it's really all about doing as much regulation as we need to to protect people and as little as we can to encourage innovation.Sam Hawley: Toby, what do you think? Can we actually let AI just rip, do what it's going to do, to reach its ultimate potential, whatever that might be, and we don't know, of course, at this point, and protect jobs at the same time, or do we actually just need to accept, like with revolutions in the past, that jobs have to go, but life in the end will be better and easier?Toby Walsh: Well, it's certainly the case that some jobs will go. I can't imagine that we're going to have so many truck drivers when we've got self-driving trucks. I can't imagine we're going to have so many people in call centres answering calls when computers can pick those calls up at 3 o'clock in the morning and do it so much cheaper. But equally, I think we have to worry about, how do we support those people? How do we ensure that those people find new employment? What skills are they going to need? How are they going to get those skills? That's the role of government, as well as the individual, as well as the corporation, to help people through those difficult transitions. What's the unique challenge here compared to previous transformations, general-purpose technologies that transformed our lives, is the speed with which it's happening. We've never had a technology that's ended a life so quickly. The Industrial Revolution took 50-odd years to happen. This one seems to be happening overnight because we've already put the plumbing in, and you can take a new AI and you can reach a new audience of a billion people almost overnight. We've never had technologies that we could so quickly, so easily, so cheaply put in the hands of so many people. And it does seem that we're living through exponential times, and that, I think, poses the unique challenge here, that we don't have the time to adjust. It's not just our children who are going to have to deal with it. We're going to have to deal with it as well.Sam Hawley: Toby Walsh is the Chief Scientist at the AI Institute at the Uni of New South Wales.