Latest / The Morning Edition / Guardrails or ‘world leader’: How Anthony Albanese responds to OpenAI’s hack

Transcript

Guardrails or ‘world leader’: How Anthony Albanese responds to OpenAI’s hack · The Morning Edition

0:03
S1: I'm Stephanie Peatling, filling in for Jacqueline Maley, and you're
0:06
S1: listening to Inside Politics from the Age and The Sydney
0:09
S1: Morning Herald. Today we're going international to talk about Prime
0:16
S1: Minister Anthony Albanese trip to New York for the annual
0:19
S1: meeting of the United Nations General Assembly. It's been quite
0:22
S1: an eventful few days for the PM, who has, among
0:25
S1: other things, spoken about Australia's attempts to crack down on
0:28
S1: social media, met with US President Donald Trump and given
0:32
S1: a speech in which he declared the post-World war II
0:34
S1: World Order no longer fit for purpose. But all of
0:38
S1: that was nearly forgotten when Mr. Albanese stood up in
0:41
S1: what was very early on Thursday morning, our time to
0:44
S1: announce an open AI agent infiltrating a Medicare website in June,
0:48
S1: accessing both public and private files. Joining me in Canberra
0:53
S1: is chief political correspondent Paul Chuckle and live from New York,
0:57
S1: where he is travelling with the Prime Minister, national correspondent
1:00
S1: Rob Harris. Let's get started. Let's start with you, Rob.
1:06
S1: You've been travelling with the Prime Minister since he left
1:08
S1: Canberra last week. And this morning he had some pretty
1:10
S1: surprising news to announce. Can you talk us through what happened?
1:13
S2: Yeah, well, with with not much notice. We got invited
1:17
S2: to the consulate here in New York where the PM
1:21
S2: came out and announced there'd been sort of a major
1:23
S2: security breach, in his view, which was an AI agent
1:28
S2: accessing a Medicare linked site. We don't know a hell
1:34
S2: of a lot. Well, we know a bit. We know
1:36
S2: that it's no private information so far. But in the
1:40
S2: scheme of things, and in the scheme of, I guess,
1:42
S2: AI debate at the moment, it's a, I guess, a
1:46
S2: pretty important reminder of the power behind this stuff.
1:49
S1: The incident itself is disturbing enough, Rob, but there's also
1:51
S1: a very murky timeline here, which raises some very difficult
1:55
S1: questions for the government. What happened and why did it
1:58
S1: take so long for the Prime Minister to tell people
2:00
S1: what took place?
2:01
S2: Well, that's a very good question. And it's not only
2:05
S2: the government, but I guess it's it's AI as well.
2:07
S2: So what we know is that around about June 18th,
2:10
S2: the open AI agent was basically sort of scraping a
2:15
S2: Medicare statistics site and penetrated further than perhaps it should have. Now,
2:21
S2: open AI found out quite a few weeks later that
2:24
S2: this has occurred. And from what we can gather today
2:27
S2: on around about September 10th. So that's three or so
2:31
S2: months afterwards, it notified the government. Now, the Prime Minister
2:34
S2: said it sent an email to sort of a public
2:36
S2: email address, and that's how it first found out. Now,
2:40
S2: the Prime Minister was advised on the weekend. So it
2:44
S2: sort of means it's what's the 24th or 25th today, um,
2:49
S2: depending on what time zone you're in. So that means
2:52
S2: 15 days since the government first found out or since
2:55
S2: government agencies first found out and about four days since
2:59
S2: the Prime Minister found out. Now they are saying that ASD,
3:03
S2: the Australian Signals Directorate, which of course is the cyber
3:06
S2: authority in Australia where we're checking things and investigating. And
3:12
S2: that was the holdup between the Prime Minister being advised
3:15
S2: and and the announcement today. Now, the PM said, you know,
3:18
S2: he wanted to be open and transparent about this. This
3:20
S2: is a potential sort of threat to people's data and
3:24
S2: very personal data, obviously, within within Medicare. And that's why
3:28
S2: I wanted to come clean. And we have also found
3:31
S2: out three other sites, a statistics site, the Australian um
3:35
S2: health linked one and a couple of state ones are
3:38
S2: potentially also involved. But yeah, it's, it was very politically
3:42
S2: convenient for the PM. Now you can be cynical of this.
3:45
S2: We've got no reason to doubt that they deliberately held
3:47
S2: this up to use it for political advantages. But if
3:50
S2: you wanted to, it would also be a perfect scenario. So, um,
3:54
S2: we'll just I think people will have their own views
3:56
S2: about this. Steph.
3:59
S1: Far be it from us to have conspiracy theories. Rob. Paul,
4:02
S1: let's bring you in here. While Rob was in New
4:04
S1: York with the PM. You were watching Deputy Prime Minister
4:07
S1: Richard Marles and Government Services Minister Katy Gallagher try to
4:11
S1: walk us through the timeline of what happened. What extra
4:14
S1: information did they give us that can shed light on
4:17
S1: the sequence of events and the severity of the actual incident?
4:21
S3: Well, in terms of the severity, to Rob's point there,
4:24
S3: in terms of how significant an event this was, there's
4:28
S3: no doubt that Richard Marles, who's the deputy prime minister
4:30
S3: and has been one of the AI spruikers inside the cabinet,
4:33
S3: used slightly toned down rhetoric compared to the Prime Minister,
4:36
S3: who stood up a few hours earlier in New York.
4:38
S3: At the press conference Rob was dutifully attending. Richard Marles
4:42
S3: was emphasising the point that while this was a significant
4:45
S3: event and spoke to broader risks, particularly around recursive AI,
4:50
S3: which are these models that teach themselves and might break
4:53
S3: out of human control, which has created a significant amount
4:56
S3: of concern even among the AI billionaire owners. He did
5:00
S3: add to that that this was a relatively minor incident,
5:03
S3: that there was no personal private data breached. Now, having
5:08
S3: said that, he does say that, you know, our more
5:12
S3: significant national security information has a fortress in front of it,
5:16
S3: and we're confident that AI models would not be able
5:19
S3: to creep around it, but that this information, even though
5:23
S3: it was not hugely sensitive, was not all public. And
5:26
S3: so there was a fence that needed to be jumped here.
5:29
S3: And so the government's message is that, yes, this was
5:31
S3: not a significant privacy breach, but it still represents a
5:35
S3: cyber attack. That's the language that Andrew Charlton used today
5:38
S3: on radio. And so that it does speak to a
5:40
S3: broader risk. Now, to Rob's point. There's no doubt that
5:43
S3: this fits into a broader narrative that the government is
5:46
S3: trying to sell in New York, and the Prime Minister
5:48
S3: is trying to sell around digital safety. It's trying to
5:51
S3: be a world leader. The Prime Minister sees it as
5:52
S3: a part of his political legacy. Now digital duty of care,
5:55
S3: the social media ban for teens and now, in his mind,
5:58
S3: lead in the world on AI. But people could make
6:02
S3: up their own minds as to whether the Prime Minister
6:03
S3: might have used slightly toned down rhetoric to make the
6:06
S3: points that Richard Marles made later on about the severity.
6:10
S1: AI has obviously been a huge policy issue for the
6:12
S1: federal government this year, Paul, as it has for governments
6:15
S1: around the world. And it's using phrases like establishing guardrails
6:19
S1: and making sure humans are in charge, while also using
6:22
S1: terms like fortress and fences and so forth. But it
6:25
S1: certainly doesn't feel like the humans are in charge, does it?
6:28
S3: Yeah. Well, in, in this conversation now, which has been
6:31
S3: bubbling away for years but has really heated up in
6:34
S3: recent months, it's AI and guardrails and safety have become
6:37
S3: almost the diplomatic and political buzzwords. And Rob would be
6:40
S3: experiencing that in New York. But in in all of
6:43
S3: this conversation, there is a sense that the risks are
6:47
S3: somewhat esoteric to a normal voter mum or dad in
6:50
S3: the street who runs a small business or is going
6:53
S3: to their public service job at most. Uses AI to
6:56
S3: put some inferences in and learn information more quickly, or
6:59
S3: to perhaps schedule their lives or to, you know, give
7:02
S3: themselves some more convenience. But this incident with Medicare data,
7:08
S3: some of which is public, some of which was not
7:10
S3: being breached, is one of the first real world examples
7:13
S3: of what these systems can do in a bigger picture sense,
7:17
S3: in terms of compromising the way government works or the
7:20
S3: way critical systems work. And even though this was not
7:23
S3: an attack from a nefarious foreign country like China, which
7:26
S3: would create even more worry, and which we do know
7:28
S3: is happening all the time, this has brought to life
7:31
S3: one of the broader risks of AI in terms of
7:33
S3: these systems breaking out, as I've said, of human control.
7:37
S3: And I was just reading some reports in the Financial
7:39
S3: Times and BBC, which has this story at the top
7:41
S3: of its website, so clearly has captured global attention. This, apparently,
7:46
S3: according to some analysts, is the first known hack of
7:49
S3: a government website by a AI model, an AI bot.
7:53
S3: So even though it is minor in some regards, and
7:57
S3: there is a, I think, plausible argument that the government
7:59
S3: has used this politically, this has become a global breakout
8:02
S3: moment to exemplify the problem that people are becoming worried about.
8:07
S1: Yeah, right on that, Paul, because we've seen, you know,
8:10
S1: multiple hacks, cyber attacks in the past few years, but
8:13
S1: they've been by people basically working behind the screens of
8:18
S1: their computers to try and get access to people's data
8:21
S1: for malicious reasons or just for the heck of it. Basically,
8:25
S1: we've seen it happen at multiple telecommunications companies. It's happened
8:29
S1: at our universities. Uh, it results in great, you know,
8:33
S1: annoyance to most people because they have to change passwords
8:36
S1: and so on. But it's a whole different level, isn't it,
8:38
S1: when it's not even an unnamed or unfaced human sitting
8:41
S1: behind a computer screen, it's, it's a bot.
8:44
S3: It is. And it's not a 25 year old hacker
8:47
S3: sitting in a bot farm in Russia or or China.
8:50
S3: It's a company who OpenAI and anthropic and others are
8:55
S3: firms that have been, you know, Australia is bending over
8:57
S3: backwards in some ways to try and lure these companies
9:00
S3: to do major investments in Australia, because, as Richard Marles
9:03
S3: has said, when he went over to visit Sam Altman
9:06
S3: only two weeks ago, as Richard Marles said at that time,
9:09
S3: this is the biggest moment in Australia's sovereignty trajectory since
9:12
S3: World War Two in terms of trying to bring ourselves
9:15
S3: into the development of these technologies, which he compares to
9:18
S3: the nuclear arms race, and needing to be close to
9:21
S3: the good guy in the nuclear arms race, being America,
9:25
S3: with which we have an alliance rather than China. So
9:27
S3: there's a perplexing kind of paradox here that will be
9:29
S3: very hard for the general voter to understand. Whereas on
9:33
S3: one hand, the government is saying that we need to
9:35
S3: become much closer to the companies developing these technologies. But
9:39
S3: on the other hand, our systems have been breached, and
9:42
S3: we need to put up some guardrails and make sure
9:44
S3: that we're able to manage the risks. So it's just
9:47
S3: another moment in this new tech world where people will
9:51
S3: be left scratching their heads and the world seems to
9:52
S3: be moving very quickly.
9:54
S1: Rob, there's been a bit of a split within the government,
9:56
S1: hasn't there, for the for the past little while now,
9:59
S1: about how hard it should go in on attempting to
10:02
S1: regulate AI or on the other side, how quickly it
10:05
S1: should embrace AI. There's been ministers like Andrew Charlton who
10:09
S1: are pretty enthusiastic about its benefits. They don't want Australia
10:12
S1: to be left behind, while other people, such as former
10:15
S1: minister Ed Husic, think that we should be much more cautious.
10:18
S1: How is this going to affect the internal conversations within
10:21
S1: the government?
10:22
S2: Well, I, I think pretty much the Prime Minister has
10:26
S2: pointed to where we are now. Look, AI is a
10:29
S2: lot of different things to a lot of different people.
10:32
S2: To some groups. It's about Australian creatives and sort of
10:35
S2: writers and music and songwriters and cultural people, and that
10:41
S2: probably appeals to the PM. I would say, you know,
10:43
S2: his his background, he loves to be loved by that crowd.
10:47
S2: And there's a big argument there about copyright and around
10:50
S2: sort of the cultural sector. But then, you know, there's
10:52
S2: issues like data centers and people who worry about data
10:56
S2: centers in their neighborhood, worry about water and data centers,
10:59
S2: worry about energy in data centers. And then there are
11:03
S2: issues like what we're seeing now, the incredible power of
11:06
S2: AI and what it could do to the potential of
11:09
S2: our security. And plus what it might do to middle
11:11
S2: class jobs. So there's a lot of different issues. And look, simply,
11:15
S2: I think it's all that, you know, where they're where
11:17
S2: they're going to fall is where they get a balance, right?
11:19
S2: I mean, you're not going to, you're not going to
11:22
S2: really pull back from this stuff now. But it is interesting. Obviously,
11:26
S2: we were part of a a group of 22 countries
11:29
S2: that at the UN this week, which involved France and
11:32
S2: Finland and Sweden, a lot of sort of OECD countries
11:36
S2: basically saying, look, let's let's be careful about this stuff.
11:39
S2: Let's put some guardrails in place. The only problem is
11:42
S2: Donald Trump here in America isn't really interested in that.
11:46
S2: He is basically only sees the economic advantages. Obviously, a
11:50
S2: lot of the AI companies are based here. He wants
11:53
S2: to win the AI race, as he says on his
11:56
S2: social media. And that race is basically between China and
12:00
S2: the US. So Donald Trump doesn't want to put much
12:03
S2: hindrance on that. Now, the interesting thing is he is
12:07
S2: really the one that's out there on his own, to
12:09
S2: be honest, because even the AI companies now, Sam Altman
12:13
S2: and the fellow firm anthropic, whose name escapes me for now,
12:16
S2: Dario Paul, you might know his name. Yes, that's the one.
12:20
S2: That's the one. They were both at the UN today,
12:23
S2: basically admitting that they need to be held back. You
12:26
S2: know that there needs to be these guardrails. So are
12:28
S2: we going to see a position? I don't think anyone
12:30
S2: will trust them to regulate themselves. And that's why they
12:33
S2: want government sort of to intervene. But as we see
12:37
S2: with sort of social media bans and digital duty of care,
12:40
S2: the US doesn't really want foreign governments putting any regulation
12:44
S2: on this stuff. so I suspect we'll find a way
12:48
S2: through it. But until we do, it will be a
12:50
S2: really pointy political debate.
12:52
S3: And worth noting that amid this debate, there is a meeting,
12:56
S3: as we're speaking, between the two men who have most
12:59
S3: the greatest ability to shape this in the president of
13:01
S3: China and the president of the US. And as you say, Rob,
13:04
S3: there's this kind of shadow boxing and posturing going on
13:07
S3: between the leaders of the tech firms themselves, with anthropic
13:10
S3: and OpenAI saying we need regulation but not doing much self-regulation.
13:15
S3: On the other hand, there's Mark Zuckerberg and Jensen Huang
13:17
S3: of Nvidia, who this week have said it should be
13:20
S3: on the firms themselves. There should be no overarching framework
13:23
S3: to regulate. And on the Trump posturing on how he's
13:27
S3: been speaking about regulation and pushing firmly against it to
13:31
S3: try and get ahead of China. It was only in
13:34
S3: June this year, or perhaps July, when the Trump administration
13:38
S3: and the white House. This apparently went all the way
13:40
S3: up to Trump, pulled back on Anthropic's latest model because
13:43
S3: of its risks and said, we can't be sending this
13:46
S3: across the world. And the white House took a very
13:48
S3: hands on role, which was a real sharp turn for Trump.
13:51
S3: And that prompted a big shift inside the Australian cabinet.
13:54
S3: There was a cabinet discussion on that, and the view
13:56
S3: at that time in Australia was, Holy crap, Trump might
13:59
S3: actually take a hands on role here. We don't know
14:02
S3: when we'll be cut off from these models, which might
14:05
S3: become crucial to our defence and national security and economy.
14:08
S3: And we need to do something to link ourselves into
14:11
S3: the US system so that we can create leverage within
14:14
S3: it and become closer to the firms and bring them
14:16
S3: into Australia. So there's all these competing tensions, and the
14:20
S3: conversation is moving so quickly that I don't think it's
14:23
S3: so unpredictable as to where the US even lands on
14:25
S3: this question of regulation.
14:26
S2: Well, just to that, Paul, I mean, I don't think
14:29
S2: I was even on the agenda at the UN this
14:32
S2: time last year. So this is this is where the
14:35
S2: debate has gone so quickly. I know journalists weren't writing
14:39
S2: about this stuff 12 months ago. I mean, this is
14:41
S2: a this is something evolving so quickly, so fast that,
14:44
S2: you know, I don't know where we'll be tomorrow, let
14:46
S2: alone next week, let alone next month, let alone next year.
14:48
S3: Totally. And the Prime Minister wants to position himself as
14:51
S3: a world leader on this. Some might portray that as cynical,
14:54
S3: but in terms of him selling his message back home
14:56
S3: as to why he's over in the UN talking about
14:58
S3: tech as a small a leader of a pretty small nation,
15:01
S3: this incident does add weight to his argument to Middle
15:05
S3: Australia that this is worth this is worth talking about
15:08
S3: on the national stage.
15:09
S1: On that point, though. And Rob, let's chat to you
15:11
S1: about this for a minute. I mean, as you rightly
15:14
S1: point out, at this meeting last year, the Prime Minister
15:17
S1: was in attendance to talk about what was Australia's crackdown
15:22
S1: on social media, which is kind of a similar story
15:24
S1: to what we're seeing with AI. Social media many years
15:27
S1: ago was seen as this exciting new technology that would
15:31
S1: benefit us all. Lo and behold, a few years later,
15:33
S1: we all started to think a little bit. You know,
15:35
S1: more about that and decided maybe it wasn't so good
15:37
S1: and tried to regulate it. That's proved quite difficult. And
15:40
S1: yet the Prime Minister has continued to be out there
15:43
S1: with attempts to regulate it. Now he's back at the
15:46
S1: same meeting with a similar message on AI. Is he
15:49
S1: a world leader on this, Rob?
15:50
S2: Look, I think they bring it back to that social
15:52
S2: media point quite a lot. They say, look, it took
15:55
S2: us a long time. It took us a long time
15:57
S2: to realize and perhaps it was almost too late in
16:00
S2: many cases that, you know, a lot of young kids, obviously,
16:03
S2: we've seen a lot of young kids lives have been
16:05
S2: changed by social media. And perhaps people do lament that
16:08
S2: they were too slow to affect change, and that's why
16:11
S2: they're getting out ahead on this. Look, I mean, I've
16:14
S2: obviously been cynical earlier on in this podcast about the announcement.
16:19
S2: I think what they're doing is very noble. And I think,
16:21
S2: you know, they are showing genuine leadership. They are working
16:25
S2: with a lot of middle powers, you know, Canada and Britain,
16:28
S2: but also some of the European countries who are really
16:31
S2: interested in this and they're trying to affect change. And
16:33
S2: I think we should be very proud of that, that
16:36
S2: they're on the world stage. It is easy to sort
16:38
S2: of sneer and say they're off at the UN and
16:41
S2: talk about cost of living. But this is a real
16:43
S2: world issue. It's just a matter of fact of balancing everything.
16:46
S2: You know, unfortunately, our politics is so binary nowadays. It's
16:49
S2: like you can only be doing one thing or the other,
16:51
S2: but you should be able to walk and chew gum.
16:53
S2: So it is very good. They are playing a leading
16:55
S2: role here, whether they affect change, I'm not sure, but
16:58
S2: I think the Prime Minister made a pretty important speech
17:00
S2: tonight saying that we've got to do it together. No
17:03
S2: one can do it by themselves. And the more the
17:05
S2: more the merrier in that sense. And, uh, they've I
17:08
S2: think they've made inroads. You know, this little incident, uh,
17:11
S2: won't probably be remembered in a few weeks, let alone
17:14
S2: a few months. I wouldn't have thought. But the bigger
17:16
S2: issue remains that they need to get this stuff under control,
17:19
S2: and hopefully it's a step towards doing so.
17:21
S3: And can it even be brought under control, is the
17:23
S3: bigger question.
17:24
S1: Well, indeed. And let's maybe have one final question on
17:27
S1: that point, Paul. We've heard a lot from various people
17:30
S1: with hot takes on this this morning, but one of
17:32
S1: them is is a very knowledgeable person, former cyber security
17:36
S1: Alastair MacGibbon. He seemed pretty appalled that this had been
17:40
S1: able to to take place, and was quite shocked at
17:43
S1: the timeline and the lack of detection capability that Australia
17:47
S1: seemed to have. Is he right to be so shocked
17:50
S1: and can we expect to see a lot more investment
17:52
S1: in this area from the federal government?
17:53
S3: For sure. Well, Andrew Charlton, who's really the AI zealot
17:57
S3: in the government, comes from the tech world, has a
17:59
S3: business background, said on radio this morning that in very
18:02
S3: clear terms, this is a cyber attack and we should
18:04
S3: be treating it as such. And he said that we've
18:07
S3: learnt lessons around where our information is stored. The government
18:10
S3: has announced that they have moved some of this information
18:12
S3: behind a different kind of firewall. And so even though
18:15
S3: it wasn't personal sensitive information, it has highlighted a broader issue.
18:19
S3: And one could plausibly argue that the government's been caught
18:21
S3: on the back foot and has blamed a firm for
18:25
S3: their own failings in managing data. I think this will
18:29
S3: add impetus to the government's legislative agenda, which they plan
18:32
S3: to introduce to the Parliament early next year, which goes
18:35
S3: to AI safety and goes to also data centers and planning,
18:40
S3: but you're hearing noises from the government now that this
18:42
S3: will put a greater focus, perhaps greater speed around ensuring
18:46
S3: that data is managed in a proper way. But the
18:49
S3: way in which it permeates the broader AI debate in
18:52
S3: terms of adoption and the attraction of investment for data centers,
18:56
S3: it's hard to see this being a moment in Australian
18:59
S3: discourse that creates increased sentiment for a speedy AI transition.
19:05
S3: You'd suspect it goes in the other direction, but we'll see.
19:08
S2: Before the census, the Auditor General's office put out a
19:11
S2: report that ten years after a major cyber attack, that
19:14
S2: the website that does the census still wasn't ready, that
19:17
S2: they hadn't paid attention to enough recommendations from cyber security
19:21
S2: agencies within the government, and they hadn't invested. I mean,
19:25
S2: the reality is our politics is so cynical nowadays. There
19:28
S2: aren't a lot of votes in investing in cyber security
19:31
S2: for government websites. So it doesn't get the money and
19:33
S2: the attention it's got. So perhaps this is now where
19:36
S2: a our government's got to stand up and do the
19:38
S2: right thing instead of throwing money at the wrong places.
19:41
S3: And so the other thing to say on enforcement is
19:43
S3: that while this was not a significant breach with huge consequences,
19:48
S3: it could be a catalyst for sovereign nations to think
19:52
S3: deeply about whether further breaches and there could be many,
19:56
S3: many further breaches like this as recursive AI becomes more advanced.
20:00
S3: Whether such breaches do require strong enforcement action, and whether
20:04
S3: sovereign nations enact new laws that create clear consequences for
20:09
S3: such breaches, and the characteristic of this incident that might
20:14
S3: give greater impetus for that push is the delay in
20:18
S3: open AI. Informing the Australian Government about this incident took
20:21
S3: a couple of months. And the way in which the
20:24
S3: incident was disclosed, which was an email to a generic
20:28
S3: email based. So rather than any human to human contact,
20:31
S3: there was an email into a generic inbox from open AI,
20:35
S3: which didn't raise immediate alarm bells. It was picked up
20:38
S3: the next day and then was sent on to the
20:40
S3: Australian Signals Directorate four days later. But, uh, really, in
20:44
S3: a situation like this, you'd have an actual person from
20:47
S3: open AI make a call to an influential policymaker or
20:50
S3: bureaucrat to escalate this quite quickly. But instead you had
20:53
S3: this delay and then a really substandard email, which does
20:59
S3: go to this concern of policy makers around the way
21:03
S3: these major AI firms interact with democracies and the way
21:07
S3: they treat voter bases. You've seen this in the debate
21:11
S3: around data centers and the way they've treated some local communities.
21:14
S3: That's not across the board problem, but has particularly been
21:16
S3: acute in the US. These firms are as powerful or
21:20
S3: more powerful than most sovereign nations, and there is a
21:24
S3: growing sense among policy makers that if these AI firms
21:27
S3: don't treat democratic governments with greater respect, they will lose
21:32
S3: the social license that governments do want them to retain
21:36
S3: so that they can roll out these systems and increase
21:38
S3: productivity and national defenses. So I think there's a possibility
21:42
S3: here that this moment translates into some greater clarity around
21:46
S3: clear laws for such breaches.
21:48
S1: We can only hope, Rob, we'll let you get to bed.
21:51
S1: Thank you both for your time.
21:53
S3: See you in your back rub. Bye, Steph. Cheers.
22:00
S1: You can read all our political news on our website,
22:02
S1: theage.com.au or sm.com.au. Today's episode was produced by Josh towers
22:10
S1: with help from Debbie Harrington. Our executive producer is Tammy Mills,
22:14
S1: and our podcasts are overseen by Lisa Muxworthy and Tom McKendrick.
22:19
S1: Before you go, please follow Inside Politics and leave a
22:22
S1: review for us on Apple or Spotify. I'm Stephanie Peatling,
22:26
S1: thanks for listening.