DELANEY, Caitlin, First Assistant Secretary, National Reform and Youth Division, Department of Education MOON, Mr Brendan, First Assistant Secretary, Child Care Integrity Division and Chief AI Officer, Department of Education MORTON, Ms Madonna, First Assistant Secretary, Participation and Student Experience Division, Department of Education CHAIR: Welcome. I understand that information on parliamentary privilege and the protection of witnesses giving evidence to committees has been provided to you. I invite you to make a short opening statement, at the end of which we will start our questions and discussion. Ms Morton : We don't have any statement to make. CHAIR: Okay—straight into it. I might kick off. One of the things we're looking at, at the moment, and what we are starting to see is tech in schools, and I know the minister has been talking about a potential digital detox for our schools. The latest PISA data showed that, the more students are online, the worse their results are, and being online can result in cognitive stunting. As I said, we've heard from the minister on the digital detox. I understand that he'll be raising this issue at the next meeting of education ministers. Can you talk the committee through the PISA results and the impacts of screens in the classrooms? Caitlin Delaney : The PISA results did demonstrate that, the more screens are used, the more there is an impact on students' results. One of the key statistics that is worth noting is that four out of 10 Australian students were distracted by using digital resources, such as smartphones, apps and websites, during lessons, which was much higher than the OECD average of 28 per cent. Australian students were also twice as likely to use digital devices for schoolwork compared with the OECD average. They had an average of 2.6 hours a day using technology for learning activities, which is well above the OECD average of 1.7 hours. And PISA found more broadly that 46 per cent of 15-year-old students across OECD countries used AI chatbots weekly or more to help them learn. Another thing that came out of PISA was that some analysis demonstrated that if digital devices are used at school for a certain period of time between five and seven hours then this could actually be positive for students' learning. However, those who used digital devices for leisure for a certain period of time actually had negative results. So there are some issues that need to be worked on. This is something that a lot of countries are focused on and, as you mentioned, will be on the agenda for the EMM, with a particular focus on some of those risks around cognitive offloading that we can see when there's an over-reliance on the use of AI. CHAIR: Yes. I noticed earlier this month that New York City banned the use of AI in their primary schools. And New South Wales has announced limits on screen time, as has Victoria. Can you take the committee through what other evidence governments are relying on in making these decisions? Caitlin Delaney : Generative AI is still relatively new, and I think that as a result the evidence is still developing. But we do have things coming through PISA. There are some other randomised controlled trials happening around the world. A large-scale randomised controlled trial in Türkiye showed that students' results in maths reduced by 17 per cent when they used ChatGPT without learning safeguards. So I think there are a range of examples that are quite specific and those sorts of studies and analysis, in addition to the results we're seeing coming through these types of testing. CHAIR: Thank you. Just on the role of the National AI in Schools Taskforce, can you take the committee through an overview of the department's work on AI across early learning, schools and higher education—in particular, obviously, the role of the taskforce. Caitlin Delaney : Yes, I'm happy to start with the taskforce and give a bit of context about that. The taskforce was established back in 2023 and was put in place in reaction to the growing need for education ministers to consider the impacts and risks associated with AI. It included representation from the Australian government, states and territories, Catholic and independent education bodies and the national schooling agencies. That's ACARA—the Australian Curriculum, Assessment and Reporting Authority—AERO, AITSL and ESA. That was put in place to develop a framework. That framework was released on 1 December 2023, and all education ministers agreed to implementation of that framework. The taskforce also assisted with work on product standards which are being developed as safeguards for the use of AI products in schools. Recently, the taskforce has shifted into a community of practice. This reflects the fact that the framework is in place and those digital standards are now being piloted, so the focus is more on information sharing. It's the same people. They still meet quite regularly, but with that real focus on sharing best practice and learning from each other across jurisdictions. That's quite schools focused, but the key things underneath that were the development of the framework and then the product standards, which I'm happy to talk through in more detail. I'm happy to hand to others. Ms Morton : I'm looking after the higher education system, which is a little bit different to schools. Of course, in the university system we have the Tertiary Education Quality and Standards Agency, TEQSA, and it sets out a range of threshold standards that it asks universities to comply with. There are a number of those that are relevant to genAI. I'll go through those in a moment. Importantly, though, TEQSA has established what it's called a knowledge hub. It's kind of a repository of practice, information and resources about how providers and universities can grapple with the issues around AI. As Ms Delaney mentioned, it's not necessarily focused on prevention of learning in the higher education context—noting that schooling is a much younger cohort of children. In the higher education setting, it's much more about helping providers grapple with how they use AI to augment critical thinking and how they're building that into their curriculum—so the courses that they offer and the assessment program that they offer. We're certainly seeing that TEQSA is reporting that a number of universities are really changing their assessment practices. For example, they're going to be doing more face-to-face assessment. Remember how you used to do the old handwritten exams and you used to have the invigilators going around. Maybe I'm showing my age; I don't know. It was a very manual process. They're certainly looking at more techniques around oral presentations and group work where they can observe the learning in a much more direct way. So there's a whole tranche of work that's being done. TEQSA's published a couple of frameworks as well. Just to go back a step, unis are autonomous institutions. CHAIR: Yes, of course. Ms Morton : They accredit their own courses. They essentially set out how they will assess that course. Every course has the content and then the assessment that you can expect. For example, there is standard 1.4 of the threshold standards. That's the learning, outcomes and assessment standard. Universities and providers have to meet that standard. It also has a really close relationship with standard 1.5, which is the threshold standard around qualifications and certification. Again, providers have to meet those threshold standards, which basically attest to the fact that the student has had a high-quality learning experience, that they have learned the thing that the course says that they will learn, that they get that credential for that qualification, and that they have met all those mandatory requirements so that they can actually be a social worker—not that they've gone off and done something else. There are also other cross-cutting standards around standard 6, which is around corporate governance for institutions and how they manage risks, and AI is one of those risks that's been identified, as is quality assurance—how they're assuring the quality of the offering that they're providing to students. One more thing I will mention is that the minister mentioned at the Australian Financial Review summit—back in, I think, in mid-August, around the 18th—that he was looking at TEQSA's powers and whether or not they were still fit for purpose as a modern, contemporary regulator. He talked there about reviewing those threshold standards and bringing legislation before the parliament around that—I think they were done in 2021—just to make sure that they still reflect a modern learning environment and have that capability to adjust and give appropriate guidance to providers and that they're holding them to account so that any student who is enrolled and studying a course can be assured they're getting the value and the benefit of that and so that there's academic integrity that sits with that as well. We don't necessarily offer direct guidance to universities around that. Most of it is done through the TEQSA framework, and universities are quite used to playing in that space. They've been aware of this journey for a little while now. Going back to, I think, 2023, frameworks were starting to emerge, which probably lines up with the schools taskforce. I was working in the Victorian Department of Education when that AI taskforce was happening and I fondly remember the need for that practical experience in the classroom. Around 2023, coming out of COVID, when everybody went online, you started to see online pedagogy being a focus: how do you teach a classroom of kids through a laptop? Then that started to morph into generative AI and how you use that to augment what you're doing rather than to automate what you're doing. So it's something that universities have been working on, and it's quite pleasing to see that they are sharing information with one another. That TEQSA knowledge hub is actually a very well credentialled resource, and you'll see things on there that universities themselves have done think pieces on and are sharing with the broader community. Unless you have any questions, I might pass on to Mr Moon. CHAIR: Not at the minute. I was just going to throw to early learning. Mr Moon : From an early childhood perspective, probably the strongest thing for the taskforce to draw is that, where a preschool is associated with a school, there would be correlation so that some of the learnings and better practice being shared would relate to the preschool setting as well. In early childhood more broadly, it's less prevalent in terms of student learning on the ground, because of the age of the children and the pedagogy. I would note, though, that, from a whole-of-sector sustainability and productivity perspective, many of the things are what we see in the broader economy, noting that a significant proportion of early childhood services, particularly childcare services, are small to medium enterprises and the productivity promise of artificial intelligence is likely to apply in similar ways to those as it would to other SMEs. We are broadly aware of different use cases in the adoption of AI in larger providers for similar things, again, in terms of staff scheduling and rostering and the automation of back-office processing. CHAIR: I might leave my questions and come back to them if we have time. I want to throw to the Deputy Chair. Senator DUNIAM: Mr Hastie has a couple of questions first. CHAIR: Go for it, Mr Hastie. Mr HASTIE: Good afternoon, everyone. Thanks for your time. I'm just going to ask some simple questions about getting the nation prepared for AI. Given that our Naplan and PISA scores over the last 20 years have been on a slow downward trend, what's your view about lifting those scores—again, they're just scores—and lifting the standard of education across the board? And what areas are you planning to focus on as we try to uplift the nation in preparation for the AI revolution, which is coming whether we like it or not? Caitlin Delaney : As I mentioned earlier, there will be a discussion at the education ministers meeting around the impacts of AI. I might take a step back first of all and just make clear that the Commonwealth is responsible for providing strategic direction and national leadership around schools, but it's actually the states and territories and the non-government education authorities that make the specific decisions about how they're going to deliver and manage things like AI in their schools. We have, to date, done that by working with the states and territories on the framework and then setting up the guardrails through the product standards. All education ministers agreed, back in 2023, that there are a lot of opportunities associated with AI but there are also risks, which is something that needs to be a continued focus as AI use increases within the schooling space. That continued engagement with ministers across the country is really important—what they are seeing and what the schools are seeing in terms of the AI risks, including concepts of cognitive offloading and things like that—and continuing to discuss, as a group, how best to ensure that our children are getting the best education possible. Obviously, AI and being able to use AI will be important for children, as they make their way through to the workforce, and so it's important to find the line where kids are learning how to use it, and have the capacity to use it, but are not using it to the point that it's detracting from their capacity to learn. Mr HASTIE: I guess that's why I was asking, because of the strategic nature of your role. The question, better put, is: where do we, as a country, want our education to be in five years? Given there are risks with AI—you mentioned cognitive offloading—how are we going to mitigate those? I accept that the states and the territories obviously have the role of managing the education systems in each of their jurisdictions, but we're asking the big-picture question in this committee about what does education look like. From what I've heard today, and from what I heard when I was at UNSW recently talking to their AI team, they basically said, 'Yes, we've got to double down on the fundamentals and the basics.' It seems to me we've come full circle: once screens were the future and now we are returning to paper and pen to ensure that our kids are actually learning and internalising their education and not just offloading it. The last thing I'd say, and I just want to emphasise it—this is my view; you can respond if you like—is that moral, ethical training, the sort of stuff that is considered essential to a liberal arts education, is going to be important because we are going to need decision-makers and leaders who know how to work with AI, who understand it, who can use it in an ethical manner. That's something that we have to look at in the education system as well. I'll leave it there, Chair; thank you. CHAIR: Thanks, Mr Hastie. No response to that one? Okay. I'll throw to you, Senator Pocock. Senator DAVID POCOCK: Thank you for your time here. What other measures are you taking to ensure that students have levels of AI literacy but aren't losing critical thinking skills? Caitlin Delaney : As I mentioned before, the steps are largely—the states and territories will be working out how they do that specifically within their context and their setting. I would draw your attention to the framework that does exist. It is a high-level framework, but it articulates all of the important things that need to be considered when considering what type of AI, how much AI and all those sorts of things. The state and territory ministers have all signed up to the implementation of this framework. It is used as a means of determining their individual approaches. I can't specifically go to— Senator DAVID POCOCK: Do you know which states and territories are doing it well? Caitlin Delaney : The states and territories are all at different stages in their use of AI. It's quite mixed. It varies from states where they might make Microsoft Copilot available to teachers all the way to states that have set up their own individual AI tools. New South Wales has a tool called NSWEduChat. South Australia has an EdChat tool. Queensland have just released their tool, which is called Corella. I certainly wouldn't want to be pointing to who is doing it well. Senator DAVID POCOCK: Why is that? Caitlin Delaney : Oh, because I think these are very jurisdiction-specific issues. But I think that there is a framework in which the jurisdictions can consider how it is that they want to use AI within their jurisdictions, noting their particular contexts. Senator DAVID POCOCK: At a federal level, how do we judge—you set the guidelines here, but how do we judge what the states and territories are doing and the disparity or discrepancy between— Caitlin Delaney : Yes. There are a couple of things in that. I guess, in a way, we—sorry. Let me take a step back. We have a whole lot of national testing that looks at students' results across sectors, states and all of those sorts of things. Taking another step back, we've got the Better and Fairer Schools Agreement, which all of the states and territories have committed to. Within that, jurisdictions have committed to a whole range of quite specific things, and we are working towards a number of targets. The Commonwealth supports the approach on these sorts of things through, again, those overarching frameworks and having a look at how we're tracking against key metrics, rather than necessarily doing an analysis of how AI is used in a specific area. Senator DAVID POCOCK: Outside of the classroom, what percentage of students are using AI as a study tool or as a tutor? Caitlin Delaney : That is a good question which I might need to take on notice. Senator DAVID POCOCK: That'd be great. Are you concerned about AI chatbots and these relationships that we're hearing that young people are forming with AI and the impact that they have on them, as young people, and their schooling? Caitlin Delaney : We're certainly aware of those issues and that they're becoming more apparent. Those kinds of things will be incorporated into discussions between education ministers, going forward. Senator DAVID POCOCK: Is there any information or report you can provide for the committee to have a look at that impact? Caitlin Delaney : No, not that I'm aware of at this stage. Senator DAVID POCOCK: In terms of AI's impact on Australians' lifelong learning and the potential for a need to retrain, Independent Higher Education Australia has recommended the creation of a 'comprehensive lifelong learning framework' which potentially would include a 'portable lifelong learning entitlement' backed by income-contingent loans that workers could draw on throughout their careers for education, upskilling and reskilling. We heard earlier from the department of workplace relations that they're really not sure what AI is going to do to jobs. Is consideration being given to this sort of proposal? Ms Morton : I haven't received the proposal. Apologies—I'm not familiar with it. But we would be happy to receive it and think it through. It is fair to say, per earlier commentary, that it's something the higher education sector and the VET sector are certainly aware of and focused on. There's an interesting report—you may have already it; I apologise if you have—from Jobs and Skills Australia. It is from 2025 and about our gen AI transition. There's some really interesting data in there. What it looks at is the relationship between augmentation and automation: When are there going to be jobs in the future in which augmentation through AI is a good thing? Where jobs in the future might be automated in terms of high volume and high yield, what role would AI have to play? That report talks about only four per cent of the current workforce being in occupations that have high automation exposure to gen AI, whereas 79 per cent has low automation exposure, which I thought was quite interesting because that's really going to some of the more alarmist commentary that's out there that AI is coming for all of us, is coming for all of our jobs and will take over the world in—whenever. That was, I think, quite an important report from JSA around how it's thinking and how that will help inform government and their thinking around lifelong learning, as well as what is valued and what is invested in. On indulgence, as well, we're certainly noticing that universities are taking active steps to embrace AI. For example, we know that Swinburne university has now introduced a Bachelor of AI in Business. I think Mr Hastie has spoken with UNSW as well. They're very much launching applied degrees that bring VET, AI and academic learning together in business management, software development and cyberdefence. So you can really see them starting to respond to the future knowledge, skills and qualifications that will be needed in that context. I don't know if that's helpful, Senator? Senator DAVID POCOCK: It is. Finally, and maybe you could provide some of this to the committee on notice, are you looking at some of the emerging evidence, particularly from the US, that's looking at students and young people who use a lot of AI and the cognitive offloading, atrophy, lower brain activity, declining skills—essentially what people would call brainrot. I was reading about a study that was scanning the brains of university students who were heavily reliant on AI, and they were showing almost early dementia-type signals. Are you looking at that, and are you making recommendations to government? Because it is clearly a public health issue when it reaches that level. Ms Morton : I'll talk in general terms; I obviously can't talk in specific terms. As part of work to support government and its decision-making, we routinely look at international comparisons. On another topic, this week, I've been looking at New Zealand, Canada and the US. We certainly do look at important information that comes out of those different countries while noticing that their markets can be quite different. But they certainly give us some insights that we can have a mind to. We also supplement that with some of our own processes around—we're currently looking at the national research infrastructure road map. That's due for a refresh, which we do every five years. We're looking at the experiences of other jurisdictions around that as well, so it's something that we do routinely turn our mind to as part of our usual policy development and contemporary understanding of what's happening out there. I can't point to specific examples, but, more generally, that's certainly something that we're very tuned in to. CHAIR: Ms Mascarenhas. Ms MASCARENHAS: I want to ask some questions about the Tertiary Education Quality and Standards Agency and the role that it plays in communicating AI standards in universities, then I might ask some follow-up questions from there. Ms Morton : Sure. As I mentioned earlier, TEQSA is the body that sets threshold standards that all registered providers who access Commonwealth funding have to abide by to receive that funding. There are a number of standards that go to the issue of AI. As Ms Delaney mentioned, it's an emerging priority area. Specifically, standard 1.4, which is 'learning outcomes and assessment', requires universities to make sure that the methods that they use to assess the achievement of learning are about actual student attainment, not about some other form of attainment. There is standard 1.5, which is about qualifications and certification. That goes to how a qualification can only be awarded by the university when the requirements of the course have been set. They're two really key standards. As I mentioned earlier, the minister's announced that he will be looking at the powers of TEQSA and at the framework in which they operate and seeing whether or not those standards are still fit for purpose in a contemporary learning environment. That's a really key piece of work. TEQSA also has what it calls an AI knowledge hub, where it's got a repository with something like six elements to it. The Gen AI knowledge hub—they've got TEQSA resources, a section on academic integrity and assessment reform and a section on genAI and research. They've also got a section for student resources and support, and they've also got a government and industry bodies advice and resources section. That doesn't roll off the tongue easily. But they're certainly putting information out there to universities because, again, unis are self-autonomous institutions. They set their own courses. They set the content of those courses. They set the assessment program that goes with that, and then they take responsibility for accrediting that. If you've done a Bachelor of Social Work, you then have met those professional accreditation requirements to be a social worker. That's the framework in which its operating. There are a couple of key publications that TEQSA has made publicly available. They've got very long titles. One of the most recent ones was from June 2026 and is called Assuring quality learning in a gen AI-integrated future: The role of adaptive capabilities. That is a combination of setting out some principles for how universities can update their policies and programs and course designs as well as practical tools and tips for things that they might want to think about turning their minds to in the world of AI. Certainly, unis are getting better at sharing information with one another in terms of how they're supporting their students and their academic staff in grappling with how you use generative AI to augment that critical learning experience, not replace it. Ms MASCARENHAS: I also have watched the TEQSA YouTube videos that you've collaborated on with the University of Sydney. This is where I'll make a confession. I've gone back to university, and it's a different world because AI exists. That was one of the first videos that I watched because I wanted to understand this intersection. What was really fascinating to me was to see the way that students could potentially use AI tools but then how they can use additional tools to make it less obvious that AI has been used. There was also a discussion on how more sophisticated students who have access to more expensive AI tools might not have the fingerprints of using AI where other students might. I was wondering if there have been discussions about how that's managed. I haven't necessarily thought of the role of AI tools in, first of all, making sure that we get education outcomes but also this access and equity element. Ms Morton : I'll take your second point first. It's a really critical point. That whole digital divide is the next frontier around access, equity and participation, and I think that's relatively well understood. Some members of this committee may be familiar with Ms Leslie Lobel, who worked in the department of education in New South Wales for a very long time, and the work she has been doing with former premier Jay Weatherall around this whole thing about digital divides. You see it play out not only in AI contexts. You see it in a whole range of other areas of service delivery. It's certainly something, again, that the sector is very much aware of. You mentioned the University of Sydney. As part of that access challenge, I know that they have provided students and staff with access to AI tools for free as part of their, I suppose, offering. When they enrol with the university, they get Copilot and ChatGPT. I think La Trobe is doing something similar as well; they're partnering with OpenAI. I do know that universities are very aware of that and they don't want that to become another barrier either for achievement or for access, so it's a really great point. To your first question around how clever you can get in cloaking AI, I had dinner last weekend with my nieces and nephews who are all either studying or graduates, and the tips that they were telling me about how you can navigate your way around and through the system kind of blew me away. Senator DUNIAM: That's a mistake. Ms Morton : They know I'm an analog dinosaur, so they probably know that I wouldn't remember or know how to use it. But certainly that's something, again, that universities are alert to, and I would have to say that we are seeing signs that, as smart as you are at cloaking it, there's also another great program that can tell you how that's been cloaked as well. While there is the way to navigate around it, there's also equally being developed at the same time the detection technology that can spot that, map that and alert an assessor to that as well. I can't speak to the names of those particular programs, but I know that they are in existence, in development and also being deployed as well. So I sort of have some optimism in the sense that, as smart as we can be in using it and cloaking it, there will also be something that's the antidote to that a little bit. Mr Moon, you're a bit more knowledgeable on the topic than I am. Mr Moon : More broadly in the technology space, there are many use cases or examples where the technology itself is being used as a defensive tool—in cybersecurity, in assessment, in many regards. Smart use of the tech as a line of defence against misuse of the technology is absolutely something that's coming out more and more. CHAIR:  Senator Duniam. Senator DUNIAM: In terms of the process that we've been talking about broadly here that the Australian government is pursuing in terms of the standards for AI and data centres, has the Department of Education been asked to be a part of the work that the government's undertaking, and if so, what does that look like? Caitlin Delaney : We are represented on the Commonwealth IDC that's being run by the Department of the Prime Minister and Cabinet. Senator DUNIAM: Is that through the Office of AI or is that a separate— Ms Morton : Yes. We have, as Ms Delaney says, and we have different touchpoints for that. We'll have that through research, through schools and through teaching and learning. It is the Office of AI in PM&C. We also work quite closely with DISR, the Department of Industry, Science and Resources, who developed the national AI plan. We certainly have a seat at the table in that regard. Senator DUNIAM: What's your input to this current body of work that's going on? What does education—is it about the standards of students and what needs to be done to prepare students to come out the other end of the education system as AI literate and ready to embrace? What is the contribution? Ms Morton : As I mentioned, there are a couple of streams of work. Research is one, in particular, that we're quite active in. I mentioned the National Research Infrastructure Roadmap that's being developed at the moment—that's a five-yearly process—and that is really about where government thinks about investing in the future in terms of research infrastructure and what supports that. AI is one of those transformative capabilities. It's kind of like, if you're in the roadmap and acknowledged in the roadmap, that then sets a priority that when we look at funding, grant rounds and programs; it's mapped back to that critical piece of work. So definitely around research is one. I think knowledge, skills and education we do jointly with DEWR, so, again, it would be thinking about how AI is featured in terms of that underpinning journey around knowledge, skills, learning, probably participation and productivity, as well; that would be the other lever. When we're on an IDC, there might be ideas that are brought to the table about how we assist in the framing of putting together that plan, and then, like I said, if you're in that plan, that is sort of a strategic lever and strategic hook that guides government investment from there. Senator DUNIAM: Just going back the other way then, we talked before about what states and territories might be individually doing in their education systems. Given, as best as it can be put in this term, the national approach we're taking, is there anything beyond—you mentioned the framework, which I don't have my head around completely. But is there anything being done to try and standardise the approach taken across the country? You mentioned ministers are meeting, but is there any body of work being done so that we have a common approach across the country as to how we deal with AI in schools? Caitlin Delaney : Outside of the framework that exists, I think— Senator DUNIAM: When was that set up, again, sorry? 2020, was it? Caitlin Delaney : Yes, 2023, I think. Senator DUNIAM: So, it's three years ago, and the world is now very different. Caitlin Delaney : The world has moved on. Senator DUNIAM: It'll be different in a month's time with this emerging technology. I'm just wondering— Caitlin Delaney : I— Senator DUNIAM: Yes, go on, sorry. Caitlin Delaney : All I was going to say is that will be one of the key features of ongoing discussions between education ministers. Ms Morton : That's why they're perhaps revisiting it in November, given it feels like there's been a bit of a sharp ascension of AI issues over the last nine months or so. It's really come into play, and, certainly, the minister has been very strong on the digital detox agenda and how we are— Senator DUNIAM: Yes, it's getting along. Ms Morton : Yes. For something of this nationally significant priority, it's appropriate for the minister to speak with his colleagues about how we best tackle that challenge, that intersection between frontline practice in the classroom and then the strategic levers, as Caitlin Delaney has pointed out, that the government has. You need that sweet spot. Senator DUNIAM: This is my final question on that. I know there is only so much you'll be able to say in response to this, but would it be the case that the Commonwealth minister, Minister Clare, would go to that meeting with a proposal around how to best structure that, or is it an open conversation? How would that normally happen? Mr Moon : In a variety of ways. It would not be uncommon for any given education minister to bring forward a paper. Senator DUNIAM: Under their own name. Mr Moon : Yes, sure. Senator DUNIAM: I guess more specifically: is the Commonwealth taking a paper to that ministerial meeting as to how to deal with that? If you can't answer that, take it on notice, perhaps. Ms Morton : I think that's probably one we would be circumspect about. I'm happy to take it on notice. Senator DUNIAM: Yes, if you could. This is only a good-faith question. Ms Morton : No, we understand. Senator DUNIAM: Because this is all interrelated and I'm just hoping to see some— Ms Morton : It could be that states and territories may want to bring things forward themselves, so we wouldn't want to pre-empt that that may happen. Senator DUNIAM: Indeed, yes. But if you could take that on notice, that would be helpful. Thanks, Chair. CHAIR: Senator Sheldon. Senator SHELDON: The use of generative AI in research raises some risks relating to fabricated references, hallucination of data, inappropriate manipulation of research et cetera. There is a long list. There is also misattribution of authorship and integrity of the scholarly record. Given that the Research Council sits within the department's portfolio, there are questions about how these risks are being monitored and managed across the research sector. For the ARC, as it's under the department's portfolio, what does the department take into account when assessing the risks to Australian research quality and the research sector, and protection from AI slop entering published journals? Ms Morton : It's a really good question. Certainly, the ARC is an independent body. We don't direct how it conducts itself. I may have to take aspects of that on notice in terms of what it's specifically doing. Certainly, there are a number of processes that it uses to monitor the risks in the system. As a bit of a parallel, they are national security matters and sovereignty issues. They are very used to identifying risks and managing those risks. Sorry, Senator, I will have to come back to you on notice with the detail of that. It is not my current remit. Senator SHELDON: That's okay. I'll put it on notice and send it over to you properly. Ms Morton : Thank you. It's a great question. Senator SHELDON: There's that thing about getting that right. Then there's also a question about how it's a benchmark as well for everything else, to a degree, I think, to be tested against. There may be a good reason not to do something for a particular reason, but there's a benchmark there. Ms Morton : The other element that I would put into that mix is that in research you get much stronger peer review. You see researchers looking at other researchers and not being backwards in calling out anything dodgy. There is a market moderation mechanism that shouldn't be operating in and of itself. As the technology develops, there will need to be guardrails. I'm happy to take that first part on notice. CHAIR: Thank you very much for your appearance before us today. There are a couple of questions on notice there; could those responses be provided to us by 25 September 2026?.