- Opportunities as educational tool
Educational uses
4.1Generative artificial intelligence (GenAI) can be used in educational contexts in many ways. The technology can act as a standalone tool or be integrated into other systems and platforms. GenAI presents ample potential opportunities for students, educators, and the broader education workforce. As shown by Figure 4.1, GenAI tools such as ChatGPT can perform a wide range of tasks to assist users in the education space.
Figure 4.1Uses of ChatGPT
Source: Source: UNESCO, ChatGPT and artificial intelligence in higher education: quick start guide, p. 9.
4.2Many educational institutions are already using GenAI in ways identified by submitters:
- as a possibility engine to generate alternative ways of expressing an idea or rewording information, which increases accessibility for students at different learning stages
- as a Socratic opponent or debating partner to help students develop an argument, filter for bias, and refine their critical thinking skills
- as a study buddy to create or prompt questions, personalised quizzes, and reflection for students, especially as a revision tool. The artificial intelligence (AI) study buddy can also evaluate a student's understanding and retention of knowledge
- as a personal tutor for students to give them immediate feedback on their progress, helping them to identify areas for improvement. This could especially assist students from disadvantaged backgrounds if they need to improve their skills and do not have equitable access to teachers
- as a co-designer to assist in the design process. For instance, it could summarise research or concepts, and input it into lesson plans.
- An early adopter of GenAI in education is Pymble Ladies’ College (PLC). In its submission, PLC outlined ways that GenAI could potentially support students, including as:
- ideation partner: can assist students in brainstorming ideas for projects, essays, and other assessments
- summariser: can summarise large amounts of information into digestible chunks, which aids study and revision
- synthesiser: can synthesise information across various resources and subjects, assisting with deeper understanding and cross-disciplinary learning
- translator: can assist in learning languages by providing instant translations and language practice opportunities
- research assistant: can help with researching topics by sifting through large amounts of information and presenting relevant data
- personal guide: can provide personalised learning paths, suggest resources, and guide students through complex problems based on their unique learning profiles
- reflective companion: can encourage students to reflect on their learning, identify their strengths and weaknesses, and set achievable goals for improvement
- accessibility aid: can support students with disabilities through features such as voice-to-text transcription, text-to-voice reading, personalised learning paths, and more
- learning style identifier: can help identify a student's learning style and suggest resources and strategies that align with that style
- skill development coach: can provide exercises and feedback to help students develop specific skills, such as critical thinking or creativity
- social-emotional learning aid: can be used about understanding and managing emotions
- time management aid: can assist students in managing their time effectively by helping to plan study schedules, reminding about deadlines
- One prominent use highlighted by stakeholders was the use of GenAI as a study buddy or virtual tutor for students. Online Education Services stated that GenAI tools and capabilities will revolutionise and transform the operations and practice of the Australian education sector. GenAI could improve student experience by offering ‘adaptive tutors’ which ‘can be trained to support students using the Socratic style’. The systems can identify areas where students are struggling and offer targeted assistance and foster independent learning and improvement. As a virtual tutor, GenAI can:
- prompt inspiration, imagination, and creativity
- enable knowledge discovery or explore various topics
- simplify or expand challenging concepts
- present information in different ways or languages to encourage better understanding
- provide feedback and suggest ways to improve their own work
- The Association of Heads of Independent Schools of Australia (AHISA) conducted a survey on GenAI use in independent schools, which highlighted findings about the uses and benefits of AI. While a slender majority of its members considered it too early to tell, 43 per cent of respondents had ‘identified positive gains in either student engagement or learning outcomes, or both’. The AHISA listed these gains as:
- improvements in drafting, creative inputs, brainstorming in creative work, generating ideas
- assistance for students in research
- improvements in the calibre of students’ work
- greater understanding of concepts
- gains for students with literacy difficulties
- improvement in student engagement.
- Figure 4.2 illustrates AHISA’s survey results on the ways students are using GenAI.
Source: AHISA submission, p. 17.
4.7The findings by AHISA highlighted that the most common uses were to support student research (77 per cent), generate ideas for creative projects (68 per cent), and offer feedback to improve written text (58 per cent). Students were less likely to use GenAI to generate video, audio, or animation.
Generating ideas and content
4.8GenAI can generate summaries of long texts, rapidly identify relevant text, and ‘perform menial tasks, such as sorting information under headings or in tables’. GenAI can therefore significantly reduce students’ required study time.Students are using GenAI interfaces to generate ideas for wide purposes, ranging from consolidating concepts to getting answers to questions they are uncomfortable to ask their teachers.
4.9The Australian Science and Mathematics School highlighted how GenAI can assist students by generating ideas. It stated that students ‘having a creative block or who are stuck on choosing an idea can generate lists of ideas’, helping them to be inspired by other ideas.RMIT Blockchain similarly observed that GenAI can help the author overcome their block rather than assume the role of the writer.
4.10GenAI may also assist students who struggle with, or are anxious about, their writing skills or styles of writing. The University of Sydney encouraged educators to promote students’ use of GenAI in their work to ‘draft some introductory lines, topic sentences, or other parts of the written work to get them started while explicitly highlighting its limitations’.
4.11Tech for Social Good (TFSG)—a youth-run organisation that helps young Australians create change through responsible technology—provided examples. For instance, a student having difficulty with a first draft of their story could ‘test their ideas with ChatGPT, which can provide feedback, help draft a narrative framework and a more fleshed out draft’. TFSG claim that such an approach can assist students in testing their ideas and being more ‘confident with the creative process’.
Data-driven insights
4.12GenAI can create data-driven insights and consequently administrative efficiencies at an individual and systemic level. The Committee heard that GenAI can analyse student data, providing educational institutions with valuable insights into student progress and trends. This can be used by teachers and school leaders to measure real time student progress, to ‘understand patterns in student performance’, and to ‘predict student outcomes based on historical data’.
4.13The University of South Australia asserted that data can enable educators to ‘identify learning gaps, and provide targeted interventions, resulting in improved student engagement and learning outcomes’. The Australian Council of State School Organisations (ACSSO) stated that by analysing extensive educational data, schools can identify patterns and gaps that will allow them to make ‘informed decisions and design targeted interventions’ that will assist in curriculum development and improve educational outcomes for students.
4.14Similarly, the Amazon Web Services (AWS) explained that:
[GenAI] can generate more comprehensive feedback and guidance to teachers based on data collected from student interactions, better predict student performance, and provide personalised learning plans… An effectively trained system can generate new insights from existing data, enabling faster and better data-driven decision making (particularly with regard to student support and learner experiences).
4.15Professor Ian Reid, Fellow at the Australian Academy of Science, stated that in using AI we ‘could be analysing people's study plans, looking at their assessments and creating specific assessments or specific remedial work for students on the basis of the things that they don't understand’.
4.16Data analysis may also provide enhanced assessment methodologies. This could help to accurately evaluate student progress, streamline the assessment process, and identify additional support for students. The Curtin Student Guild observed that universities can also ‘analyse data from course assessments to gain insight into students learning patterns and needs and use these insights to assist students’.
4.17Researchers can use data-driven insights to guide them ‘through complex ideas and methods’, ‘provide bespoke tutorial content’, and ‘imitate a peer review process from multiple perspectives’. Given efficiencies like these, AWS noted that users of GenAI are already experiencing increased productivity in their learning spaces, from education businesses to front offices, staffrooms, classrooms, and at home.
Personalised learning
4.18The capacity for personalised learning through GenAI presents a unique opportunity for students to progress further on their learning journeys than ever before. The technology can assist students and teachers to focus on ‘personalised instruction, guidance and mentorship’. The Tertiary Education Quality and Standards Agency (TEQSA) asserted:
Through intelligent algorithms and appropriate data input, individuals can access a vast array of resources, courses, and knowledge, tailored to their specific needs and interests, enabling continuous learning throughout their lives.
4.19Students who have access to GenAI tools can ask them to explain concepts in multiple styles and build their understanding. The use of GenAI to explain concepts functions, similarly to an internet search, but is far less time-consuming, and further reduces reliance on teacher-led explanations
4.20Additionally, TFSG contended that GenAI can be used to personalise learning for students and to promote their creative and critical thinking skills. Students can use GenAI to ‘generate content and interactions that are attune to student needs, adapting based on the student prompts and feedback’. This one-on-one tutoring tool can provide personalised experiences, and better engage otherwise disengaged students. Similarly, GenAI can be used a ‘collaborative tool or resource bank to craft their own original pieces’ through art, text, music or code, as provided in the following example:
A student struggling to understand a teacher’s analysis of a Shakespeare play can turn to a tool such as ChatGPT to explain it in a different way or in simple terms, with different analogies or manners of explanation until the student finally understands.
4.21GenAI can be used to design personalised lessons which best match the student’s level of knowledge. This can yield positive impacts on ‘students’ perceptions of their own capability’ and therefore increase their motivation and engagement. Thepersonalised features of GenAI underpin the student’s experience and attainment of skills and knowledge from the course curriculum.
4.22GenAI can also provide personalised feedback to students, allowing them to progress their learning objectives. The ACSSO asserted that GenAI can ‘provide instant feedback and assessment to students’ and analyse and evaluate student responses on written tasks such as essays and provide constructive feedback on their weaker areas. When the technology understands a student’s strengths and weaknesses and learning styles, it can generate customised content and assessments for them to learn from. This personalised feedback can be ‘particularly powerful’ for students who have struggled with traditional learning models or have additional learning needs.
4.23Professor Reid highlighted that personalised tutoring could also be especially useful where there is a teacher shortage. Professor Reid stated that personalised tutoring could be used in regional high schools where students may have very different abilities, and in the tertiary sector, for instance, as an extension tool.
4.24One example of a GenAI tool that can provide personalised feedback at scale is the Kahn Academy’s Khanmigo tool. This tool can personalise student learning and cater to each students’ unique learning styles and needs. The tool does this by generating ‘customised content, lessons, and assessments’ that cater to a student’s strengths and weaknesses. The tool also analyses student data and preferred learning styles and suggests matching strategies, creates individual learning plans that address student’s strengths and weaknesses, and recommends resources and activities that match a student’s interests and goals.
Alleviating teacher workloads
4.25GenAI has the potential to alleviate teacher burnout by dramatically reducing administrative work and assisting with a variety of tasks. GenAI can assist educators from primary education to higher education (HE) with ‘lesson planning, curriculum design, diagnosis of student learning, and assessment and reporting’ as well as ‘grading, attendance management, and scheduling’.
4.26Independent Schools Australia asserted that GenAI tools could reduce teacher workload and increase efficiencies in the following areas:
- Identifying students who need additional support and extension, and designing intervention programs to improve student outcomes;
- Assisting in developing assessments, marking, grading, lesson planning, and generating student feedback with consistency, objectivity and fairness in grading;
- Collecting and analysing student data at scale (and low cost) and generating reports;
- Undertaking and streamlining administrative tasks such as tracking attendance, and other record keeping requirements;
- Using advanced software that can detect plagiarism;
- Determining professional learning needs and recommending resources and further learning and
- Developing a skills matrix to identify explicit skills that need to be taught to educators and students so that they can use and manage AI generated resources effectively by recognising dissonance and recognising and testing assumptions.
Administration
4.27Teachers are, on average, working 50–60 hours per week and have limited time for direct, face-to-face instruction. Access to GenAI tools is currently free and can assist teachers. The uptake of GenAI by teachers has been strong and has helped alleviate teacher workload. The Australian Education Union claimed that GenAI can reduce some teacher responsibilities, which may alleviate teacher shortages and workloads, freeing up time for more direct student engagement.
4.28Similarly, Tech Council of Australia asserted that if it is properly managed, GenAI has the potential to:
‘streamline time-consuming administrative tasks for educators to free up time and attention to dedicate to the most impactful and meaningful aspects of teaching [and] support educators to identify and adopt best-practice teaching approaches’.
4.29The use of GenAI to perform administrative tasks can benefit researchers through the collection, processing and analysis of data. GenAI can also assist with time consuming administrative and reporting processes, which could give them more time to conduct research.
4.30In school administration, GenAI is being used to rewrite text to increase its accessibility, generate question variations, build lesson plans that integrate online content, and increase marking and feedback efficiencies. As such, it can create learning resources and develop answers, and map activities to core curriculum outcomes and education prompts.
4.31The South Australian Department for Education (SA DFE) contended that GenAI can automate a variety of education administration tasks including ‘generating reports, creating content and marketing materials such as email campaigns and social media posts, creating presentations, generating ideas, and helping with brainstorming’. The SA DFE conducted a trial of EdChat to understand how it can be rolled out in classrooms, and noted that:
There was a sense that there is a need for training and professional learning to support the use of generative AI and to help understand its potential. That was flagged in phase 1 of the trial. It was used across various curriculum areas. It was used for lesson planning. It was also used as per the example I referred to earlier; it performed strongly in differentiating learning for students. Teachers, in particular, generally found that it enhanced task efficiency, which is aligned with the work we are doing in South Australia to try and reduce administrative burdens for teachers and enable greater time for teaching. It supported that objective.
4.32As Australian teachers are facing ever increasing workloads and complex tasks, there are several positive consequences of GenAI for teachers. By reducing teacher’s workload, GenAI may allow ‘more time to spend facilitating learning and helping students to develop their critical thinking, creativity, interpersonal, and metacognitive skills’. GenAI may also help provide learning interventions to support students where it is needed. The automation of tasks and subsequently efficient operational systems can increase the productivity of school administrators and principals alike, but may require more research.
4.33The Victorian Association for the Teaching of English (VATE) reported to the Committee that GenAI can assist a teacher in their daily work by offloading and reducing their workload. This raises questions about whether teachers should be undertaking tasks that can be automated by GenAI, whether the task should exist at all or if it could be done in a different way. VATE gave the example of report writing. While it can be a very onerous process, instead of offloading report writing to GenAI, education administrators could rethink the way reports are written.
4.34Reducing the administrative burden teachers and administrators face may particularly benefit schools in low socio-economic communities, where there is a higher administrative burden. The Committee heard more support is needed for families in areas of ‘poverty and trauma’. One response could be to provide GenAI resources in these communities to help manage these issues.
4.35The Australian Secondary Principals’ Association stressed the importance of providing the appropriate time and professional development opportunities for teachers to adapt to their practice to the changing curriculum. When teachers are confident and proficient in their own capabilities, they will be better placed to support students to learn technology.
Lesson planning
4.36GenAI can be used to generate educational content including lesson plans, study materials, interactive activities, and to provide topic content for their students’ year level and learning style. The SA DFE contended that GenAI can be used by educators ‘as a starting point for lesson planning’ where teachers can input the lesson objectives and other parameters and then use the platforms output as ideas for lesson design and activities.
4.37Dr James Curran, Chief Executive Officer of Grok Academy, comments on the use of AI-generated lesson plans in schools:
We've got a situation where the vast majority of our teachers are individually reinventing the wheel every night and every weekend, trying to find resources and activities. For example, in digital technologies they may have very little expertise themselves in the curriculum and therefore also not a great ability to differentiate between a quality resource and activity and something that is—let's say there are better opportunities out there. As a result, they're turning to ChatGPT, putting in a content description from the Australian curriculum and saying, 'Write me a lesson plan for this,' because we never gave them a lesson plan in the first place.
4.38The Committee heard that in many cases lessons plans that are designed using GenAI, especially for complex subjects like digital technologies, are incorrect. Dr Curran described such a situation:
We did an example at a workshop last week. We said, 'Put in this content description from digital technologies,' and it actually came up for an activity that was unrelated to that particular content description. Mostly, teachers would not have the expertise to necessarily be able to tell that that was the case.
4.39Using GenAI to write lesson plans is becoming a common occurrence, especially in HE where there is space to experiment and develop new forms of pedagogy. Universities have held training and discussions on how to incorporate course work in more reflective, interesting, and engaging ways, particularly for technical subject matter.
4.40In conjunction with lesson planning, GenAI can be used to support curriculum design. GenAI can ‘create logical sequences of learning to align syllabuses and school contexts’ and act as a type of reflective practice for teachers by ‘analysing their own teaching methods through live video capture and AI feedback’ through educational technology (EdTech) tools. Professor Leslie Loble AM, Industry Professor at the University of Technology Sydney (UTS), asserted that ‘smart’ curriculum tools use AI to bring evidence-based and ‘proven in practice’ resources directly to teachers for lesson planning’. GenAI tools that are built on ‘evidence-based pedagogy and teacher-focused support’ can provide better access to quality materials that ‘connect to required learning content and to data informed student insights’.
Grading
4.41The Productivity Commission found that ‘for every four hours a schoolteacher spends teaching, they spend one hour marking’. Furthermore, HE markers can be paid ‘piece rates that understate the time taken to mark assessments’.
4.42The University of Melbourne summarised the opportunities and of using GenAI for grading:
For example, while using automated marking (a form of AI) to provide students with rapid feedback on a low stakes multiple choice quiz might be appropriate, efficient and useful, using generative AI or automation to provide feedback on all formative assessments would likely be detrimental for students. This is because it eliminates one of the few personalised and relational interactions that students have with their teachers as well as a valuable opportunity for teachers to monitor and support student learning, their learning development, and their improvements over time. It may also embed biases, particularly where it is used to provide feedback on more subjective forms of assessment. Further, automated marking technologies operate by recognising norms and may mark down work that is unusual, meaning that they can fail exceptional work.
4.43The Committee heard that using GenAI to grade students’ assignments could save time for teachers, allowing them to redirect their focus on other aspects of teaching. GenAI-powered grading systems can ‘quickly evaluate multiple assignments and provide feedback’ to students. Grading using GenAI can be readily adopted to yes/no or multiple-choice questions as a means ‘to implement the assessment, generate marks, and provide feedback at scale and in a reasonable timeframe’. Thetechnology can assist teachers to provide more detailed and useful feedback on assessments, and ‘provide an enhanced education experience for students’.
4.44The Queensland University of Technology (QUT) asserted that if GenAI can be implemented to mark different types of assessments at scale, it would allow universities to consider other approaches to assessment that would better represent student learning. QUT stated that universities can be slow to change and therefore, policies need to be flexible and adaptive as the technology changes.
4.45The University of Sydney provided an example of how ChatGPT could be used to design and draft a marking rubric. A marking rubric is a table that helps students and assessors understand the expectations of the assignment and how a student will be marked against the assessment criteria. A good rubric will reduce teacher workload and better engage the students with the feedback they will receive. A rubric can be difficult to write from scratch and a GenAI tool can be used to generate a rubric that the assessor can finesse. The University of Sydney provided the following example prompt:
Design a marking rubric for a postgraduate assessment that asks students to apply their knowledge of the global financial crisis to a more contemporary economic challenge. The rubric needs to assess students on their use of literature, their analysis of the underlying causes of the GFC, and apply it creatively to a contemporary challenge. Please provide standards for each criterion from high distinction, distinction, credit, pass, and fail.
4.46Several submissions, however, raised concerns about the use of GenAI in grading student work. The Committee heard that using GenAI for grading is one of the higher-risk aspects of its application. It has ramifications not only for students undertaking assessments in schools and universities, but for situations when grading may affect admission to educational institutions in the first place.
4.47TEQSA contended that there is a risk of GenAI systems ‘becoming self-contained and self-referential’. If a student uses GenAI to complete an assessment and an educator uses GenAI to grade the assessment, then the ‘limited human involvement in the process undermines not just the educational experience but the very process of learning’. There is a further risk that in using GenAI to grade assessments, it will punish ‘out of the box’ and creative thinking, where only a narrow set of answers is considered ‘correct’ by the platform.
4.48The Federation of Parents and Citizens Associations of New South Wales (FPCA NSW) argued that GenAI would be incapable of assessing qualitative and substantive written work. Students could be awarded high marks for ‘writing random gibberish’ if they used ‘sophisticated vocabulary and sentence structure’. GenAI could be useful though for marking where there are definitive right and wrong answers, but it must always be verified by a person. The University of New South Wales (UNSW) states that educational institutions will need to consider where human review and feedback would be required, and under which circumstances a student could challenge a GenAI marked assessment for human review. The National Tertiary Education Union (NTEU) was concerned about accountability for decisions taken by a GenAI tool, which can affect the appeal processes.
4.49The FPCA NSW and the Centre for Digital Wellbeing (CDW) referenced the dangers of using GenAI for grading, citing the following example. Due to the COVID-19 pandemic, in 2020 the United Kingdom (UK) Government did not allow A-level exams to proceed; instead, grades were determined by an algorithm. Teachers were asked to provide an estimate of the results they expected the students to receive, and the estimations were weighted against the historic performance of individual secondary schools using an algorithm. Nearly 40 per cent of students had their grades lowered by the algorithm, with a higher representation of students coming from disadvantaged schools, while students from more affluent schools had a higher likelihood of increased grades. As the exam scores are the main criterion for entry into HE in the UK, the algorithm generated results were retracted.
VATE submitted that teachers’ professional judgement in grading may be questioned, and AI can conversely attribute higher grades to students. VATE reported a parent challenging their child’s School-Assessed Coursework (SAC) score because when entering the SAC content into ChatGPT with a prompt request for a score, the SAC scored higher in ChatGPT. This led the parent to believe that their child had been harshly graded by the teacher.
Balancing risks
4.51There are risks that the introduction of GenAI may actually increase the workload of teachers rather than reduce it. General teacher workload remains a serious concern with 75 per cent of Australian teachers reporting that their workload is unmanageable. The Independent Education Union of Australia (IEUA) argued that the adoption of GenAI needs to be viewed through the lens of the current teacher shortage.
4.52Teachers will need to learn how to use GenAI and assist students with it. There are concerns that the pace of development of these tools may outstrip the capacity for educators to keep up. GenAI may place ‘significant demands on educator time, from understanding the operations and functionality, to fully engaging with ethical and creative considerations of the tools available’.
4.53Some teachers are also concerned that their workloads will increase regarding GenAI and assessments. For example, teachers may need to run AI-detection software, fulfil authentication requirements, and double mark work to ensure education outcomes.
4.54The NTEU advised of instances where university administrations have reduced resourcing for administrative work under the guise that it can be done through Microsoft Copilot. This has not led to an improvement in productivity, rather there has been an increase in workload for teaching staff and a reduction in time spent on teaching tasks.
4.55Teachers will require adequate training to support and adapt to the use of GenAI in the classroom through pedagogical support, technical support and legal and ethical support. This is discussed more fully in Chapter 5.
Opportunities to bridge the digital divide
Access and equity challenges
4.56The digital divide presents a barrier to education equality with disadvantaged students and schools struggling to access the same digital learning opportunities as their more privileged counterparts. Many stakeholders raised possible impacts of GenAI in education as they relate to equity. Whilst GenAI has transformative capacity and the ability to improve educational outcomes, it is imperative that there are equitable opportunities for all students, regardless of their background.
4.57Australia has a significant digital divide, with one in four people being identified as ‘digitally excluded.’ Digital exclusion is ‘driven by lack of access, affordability, and digital skills.’ TFSG stated in its submission that Australia has one of the most inequitable education systems, ranking in the bottom-third of Organisation for Economic Co-operation and Development (OECD) countries in providing equitable access to education.
4.58The digital divide is demonstrated in the Australian Digital Inclusion Index. The Australian Human Rights Commission (AHRC) wrote in its submission:
The 2021 Australian Digital Inclusion Index shows that there remains a substantial digital divide in Australia. One in four people in Australia were identified as being ‘digitally excluded’ and ‘people with low levels or income, education and employment, those living in some regional areas, people aged over 65 and people with a disability’ being identified as being of particular risk of being left behind.
4.59The digital divide presents itself ‘in terms of really simple, basic technologies such as computer access, internet connectivity and data usage’. This is particularly important, as ‘generative AI requires significant digital literacy to know how to ‘converse’ with the tool, how to best prompt and interrogate, and, most importantly, how to evaluate and interpret the responses’.
4.60The digital divide is perhaps most keenly felt in low socio-economic communities, as teaching ‘institutions may have limited access to essential infrastructure to use digital technologies, such as fast internet, and may have less budget to purchase new technologies for the classroom’.
4.61The costs associated with GenAI will create barriers for some students from disadvantaged backgrounds to use the technology at school and at home. Teachers are currently trying to support students ‘who reside in tents and caravans without electricity and only a packet of chips for dinner, making learning with AI technology impossible outside of school.’
4.62Children from The Grange P–12 College in Victoria need to bring their own devices to school to use for classes and log into the school’s cloud. Teachers from The Grange P–12 College also reported that prior to the COVID-19 pandemic, the school had a one-to-one netbook per student program at its primary school campus; however, it can no longer run the program due to funding changes. The school now has laptop trolleys for one laptop to be shared between two students.
4.63This can be contrasted to many private schools, which have immediate physical access not only to GenAI tools, but also robotics and coding technology. PLC has already begun integrating GenAI into their curriculum, where they teach science, technology, engineering and mathematics (STEM) subjects from a young age and the students undertake digital technologies and data science courses where the use of GenAI is permitted.
4.64Students are aware of this divide, with one from The Grange P–12 College stating:
I think—you already have a very noticeable divide between public and private, east and west, under-represented and very prominent schools in Victoria and Australia. By not including those disadvantaged schools you're making them—it's as if you're just acknowledging they don't exist.
Regional and remote communities
4.65There is also some concern that educators in regional, rural, and remote communities may have less access to professional learning opportunities, which may impact the capacity to support digital learning in schools. In fact, the Grattan Institute has found that students living in regional and rural areas are two years behind their inner-city counterparts who attend more advantaged schools.
4.66This sentiment has been echoed by the Northern Territory (NT) Government and Charles Darwin University. The NT Government stated:
Many NT students do not have access to basic reliable technology and the ability to consistently connect to the internet remains a priority with the digital divide further disadvantaging a significant number of NT students. Innovative technology, such as generative Al, has a dependency on connectivity and digital capability of students and staff.
This is supported by a Charles Darwin University survey, which found that eightpercent of respondents in the NT were already experiencing disadvantage because they could not access the benefits of GenAI, with a further 15 per cent foreseeing disadvantage.
Access and resources
4.68Failure to address digital equity and the use of GenAI may perpetuate the digital divide in education, and Australian society more broadly. The AHRC have expressed concern that the students who stand to benefit the most form these technologies are also the most technologically disadvantaged, and least likely to gain access to it. There are significant barriers to closing the digital divide including skills accrual, cost and policy application.
4.69As GenAI is integrated into Australian schools and society, the skills needed to succeed in school and at university will change. Lack of access to the same level of technology may have long-term negative consequences on students’ outcomes. This is because addressing the digital divide requires both electronic devices and internet access, and access to high quality learning applications. The schools and families able to access the best EdTech assets will have better long-term learning outcomes, adding to their already significant advantage.
4.70An additional barrier to access is the level of knowledge that will be required to access different GenAI platforms and the ability to teach students as the technology evolves over time and becomes embedded in EdTech.
4.71Cost is perhaps the biggest and most important aspect of closing the digital divide with some cohorts struggling to access existing technology, let alone emerging technology. Data scientists are charging corporation hundreds and thousands of dollars per hour to engage in responsible AI. Disadvantaged communities may lack the ‘necessary infrastructure and resources required to access, implement and maintain relevant education technologies’. This is because of the likelihood that superior and more sophisticated versions of GenAI will be more expensive leading to accountability and quality control issues that may further disadvantage some communities. Schools vary in their capacity to pay for the additional features built into paid content, which may further exacerbate the digital divide.
4.72The digital divide is also exacerbated by the subscription-based business model that underpins AI-enabled EdTech. GenAI platforms such as ChatGPT require a paid monthly subscription, with a single account costing $240 per annum which may severely stretch the budgets of disadvantaged families and schools. Without equitable pricing arrangements for schools, it would cost $250,000 annually for a school of 1,000 students to use GenAI, which is approximately equates to three new teachers. UNSW further added that if EdTech vendors ‘insist on a pay-per-use licencing system, it will be difficult for students and universities to properly budget for GenAI use’. Professor Loble also expressed concerns that reduced cost or free versions may expose students to advertisements or weaker learning features.
4.73Students and families without internet access, devices, or knowledge and capacity will be unable to fully leverage and realise the benefits of GenAI without support. In its submission, the National Catholic Education Commission documented a recent government initiative:
A recent example is the Broadband initiative led by the NBN, which aimed to supply 30,000 broadband connections to disadvantaged families. This target was not met due to a number of factors, including perceptions of disadvantaged families and an over-rigorous application process.
4.74Resourcing disparities not only between schools, but also by State and Territory may further exacerbate the digital divide. The University of Sydney argues that the divide could widen because ‘access to and familiarisation with GenAI, because of resourcing disparities between schools and inconsistent policies pursued by different states and school systems regarding the technology’. The AHRC further claim that any framework requires ‘national consistency’, there is also a need to recognise ‘simple access to the technology… and differences in the ability of particular groups to engage with the technology’. The main danger posed is that students from more educated and affluent backgrounds will have access to much more powerful applications.
Banning the use of GenAI technology may further intensify the digital divide. Bans will not protect students or teachers from the harms of GenAI; rather, independent schools will embrace GenAI in supervised, controlled settings, leaving public and poorer schools to lag behind. It may perpetuate the already existing digital divide between those who know how to use GenAI and those who do not
Bridging the divide
4.76Although there are significant risks with GenAI exacerbating the digital divide, the technology may also help close it. GenAI can be beneficial to international students, help in the democratisation of knowledge, improve access for low socio-economic students, and help build long-term digital capacity. GenAI may help bridge the digital divide for international students. UNSW asserts that the ability for international students to use GenAI as study buddy will help them to check and edit their work before submission, ‘enabling their ideas to be judged rather than their English or grammar skills’.
4.77GenAI tools have the capacity to improve student learning outcomes and helped disadvantaged students, provided that the platforms are ‘well-designed, well-used and well-governed’. The technology has the ability to increase the accessibility of high-quality education, thereby democratising the knowledge and empowering disadvantaged students to reach their full potential.
4.78Similarly, GenAI has the potential to be transformative for disadvantaged students from low socio-economic backgrounds. By providing equal access to educational resources such as online libraries educational material and digital resources through GenAI platforms, students can have access to high quality educational content regardless of their socio-economic status, which may contribute to their long-term success.
4.79ACSSO also contends that building the digital capacity and literacy of students from disadvantaged backgrounds is critical to their future success. The ACSSO asserts that ‘[i]nteractive platforms and virtual environments powered by AI could facilitate digital skills development.’ GenAI can provide coding platforms, AI programming courses and simulations that will equip students to excel in the digital era.
4.80Stakeholders identified several ways to bridge the digital divide. These include investment and building sufficient communications infrastructure, strong standards, and good AI governance, and capacity for students and teachers.
- Investment: The GenAI market can be shaped through investment in affordable, high-quality tools that benefit disadvantaged and special needs students. There needs to be targeted resourcing for disadvantaged students and schools as well as resourcing for public schools to provide access to GenAI technology to their students. The AHRC also recommends ensuring that digital technology is available for use in community facilities such as libraries to better the digital divide. Moreover, UNSW recommends that GenAI vendors are encouraged to ‘enter into institutional licences with universities that have unlimited use for staff and students’ to ensure consistent access to similar GenAI platforms. It may also be pertinent to develop a needs-based funding model to roll out GenAI to all schools that want the technology.
- Telecommunications infrastructure: The IEUA said ‘[s]ecuring reliable internet connection and appropriate infrastructure across Australia’ will help prevent the perpetuation of inequality in Australia’s education system. Professor Loble recommended to ‘[u]tilise existing funding and regulatory leverage in the communications sector to expand access to affordable and high-capacity internet’ as:
Much of the leverage to address the digital divide will reside outside of education, for example within communications policy and investments, and those levers need to be pulled more firmly by government to overcome this basic access challenge….
- Standards and Governance: It is important to ensure that EdTech quality standards incorporate inclusive design and governance processes and that they are adhered to. CDW cautions that educational institutions should not drop their standards in order to ‘push technology into lower socioeconomic schools faster so as to be able to move that gap as this may create another layer of long-term digital disadvantage’.
- Encouraging good decision-making: Schools often know what works best for them. Public schools are unlikely to ask for millions of dollars; rather, they seek the resources that are similar to those of the school down the road and also what’s available at independent schools. Conversely, Professor Loble encourages good decision making in the purchasing of GenAI EdTech to ensure the technology is not only used, but also levels the playing field.
- Capacity building: Professor Loble stresses that it is crucial to build the capacity for teacher and students to use these tools to lower the five-year learning divide between high and low socio-economic communities. It’s not just about access to the tools, but rather having the internet access and the skills and capabilities to use them. It may be pertinent to ‘build philanthropic and other partnerships to connect digital assets—physical as well as skills and expertise—with disadvantaged schools and students.’
- GenAI could potentially support students from diverse backgrounds by improving their access to education and their learning experience. Various submissions highlighted that GenAI could assist various student cohorts, noting that more research is required to form a strong evidence base. This includes students:
- with a disability or learning difficulty who ‘require different approaches to learning in order to thrive’
- with English as a second language (ESL)
- from low socio-economic backgrounds who are experiencing equity and access issues
- living in remote or regional areas, including Aboriginal and Torres Strait Islander students
- mature age students who are often time-poor and seek to upskill.
Students with disabilities or learning difficulties
4.82The Committee received evidence about how AI-based personalised learning can assist students with disabilities or learning difficulties. For example, GenAI can enable adaptive and flexible learning environments that can accommodate students with varied challenges, such as dyslexia, dysgraphia, dyscalculia, dyspraxia, auditory processing disorder, developmental language disorder, autism spectrum disorder (ASD), or attention deficit hyperactive disorder (ADHD). For these students GenAI can provide tailored instructions and feedback that suits individual needs, preferences, and pace, and foster engagement and motivation.
4.83More specifically, GenAI can benefit students with a disability through ‘assistive technology such as speech-to-text or text-to-speech tools…language translation tools…[and] virtual teaching assistants’. Speech-to-text technology has the capacity to greatly assist accessibility for students with hearing impairments, while text-to-speech technology can assist students with vision impairments and may assist students with other reading difficulties. GenAI may assist in the development of personalised learning plans and help students with a disability to better receive instructions from adaptive AI-driven tutors, which can help improve their overall engagement and learning outcomes.
4.84The Curtin Student Guild explained that GenAI can assist students with disabilities and remove accessibility barriers:
- Image recognition can be used to create visual representations allowing visually impaired, autistic individuals or students with visual learning preferences to get a better understanding of the material;
- Brainstorming applications can assist students with ADHD to identify and structure their ideas…
- Lip reading, enhanced feedback, captioning, virtual reality and hybrid and online learning are other AI functions that can remove accessibility barriers.
- GenAI can have a strong impact on teaching and assessment, particularly for neurodiverse students. Charles Darwin University stated that GenAI will support neurodiverse students who may be very creative but struggle to adequately express their thoughts in writing. This can be achieved by providing personalised and adaptive learning experiences, tailored interventions, and data analysis for individuals with ASD. The SA DFE’s trial of GenAI has yielded such benefits. TheSA DFE noted:
The teacher set a task for the student. The student had autism spectrum disorder. They were able to use EdChat to reframe the task, in terms that were much better for the student, and also translate the task into the student's first language. It highlighted for both teacher and student that, in a matter of seconds, a task could be adapted for a student who otherwise would have had significant challenges with their learning. There were incredibly positive results.
4.86PLC agreed that personalised learning paths can accommodate a student's specific needs and abilities, but acknowledged that these technologies are constrained by their inability to provide the emotional or social support needed by some students with disabilities. PLC warned that an over-reliance on AI might make at risk students less self-reliant or hinder their ability to develop coping mechanisms.
ESL students
4.87UNSW highlighted that GenAI tools can be used by students to translate information. For example, Microsoft Azure’s OpenAI can translate over 125 languages through the Microsoft Translator Plugin. GenAI can assist by translating or correcting assessment tasks for students for whom English is not their primary language, and make a wider range of web-based information available to them. This could enhance learning for student cohorts like international students, refugees, migrants, and people raised in Australia with a primary language other than English.
4.88Professor Allie Clemans, Acting Deputy Vice-Chancellor of Education at Monash University, commented that international students were using GenAI to teach themselves and consolidate educational content rather than to complete assessments as had been assumed. However, Professor Clemans also noted that students would be less inclined to self-report if they were using GenAI to complete assessments. Similarly, Professor Nicholas Davis, Industry Professor of Emerging Technology and Co-Director of the Human Technology Institute at UTS said:
I see students with English as a second language using this, by second nature, to help them comprehend concepts and translate backwards and forwards on the fly and to give explanations of why the lecturer might have said that word. I see this at the university level, but I also hear stories from my friends and colleagues working in secondary school systems.
4.89UNSW highlighted equity considerations and how AI may be used to ‘level the playing field’ for international students who can use it to check and edit their work. AI would enable ‘their ideas to be judged rather than their English or grammar skills’. There is much to be said for permitting this practice ‘given the significance of Australia’s international student population and global educational reach’.
4.90While PLC acknowledged the constraints of artificial intelligence in translating cultural nuance and the potential for over-reliance on translations by some students, PLC highlighted how AI can ‘facilitate language practice by conversing with students’ and provide ‘cultural context and idiomatic usage of words’ for learners.
4.91The Australian Library and Information Association highlighted that ESL students are likely to feel the greatest benefits of GenAI. AI powered translation, grammar, style, and spelling tools are already used to support understanding, expression, and language acquisition. Many of these students do not have assistance at home with learning language skills, and GenAI tools can assist.
Low socio-economic communities
4.92Several stakeholders argued that GenAI can benefit students in low socio-economic communities. The OECD has found a link between socio-economic disadvantage and low academic performance. Some challenges include that students may not have access to fast internet at home, which is compounded by Australia’s cost of living issues. This can particularly affect children in out-of-home care. The Committee heard that 20percent of students in Australia in low socio-economic schools also lack access to adequate curriculum support and instructional materials. Given ‘historical challenges with equitable resource allocation’, TEQSA recommended a needs‑based distribution of GenAI with the necessary infrastructure and training to accompany it.
4.93The introduction of GenAI to the classroom stands to benefit low socio-economic students through the introduction of ‘easy, accessible, affordable and safe’ materials for student use. The Australasian Academic Integrity Network claimed that GenAI can be used to reduce this disadvantage through ‘personalised approaches that cater to different learning needs’.
4.94ACSSO contended that GenAI can ‘level the playing field for students’ through the equal provision of educational resources, stating:
AI platforms offer online libraries, educational materials, and digital resources to ensure that all students can access high quality educational content regardless of socioeconomic status. This especially benefits disadvantaged families who cannot afford costly textbooks or e-learning resources. Thanks to AI, these students can now access free or affordable educational content, which has the potential to significantly contribute to their academic success.
4.95TEQSA highlighted that GenAI tools can be used by schools to develop augmented reality experiences for students. TEQSA identified that the facilitation of virtual field trips and immersive technology will be particularly beneficial to low socio-economic schools who would otherwise not have access to the technology.
Committee comment
4.96The Committee recognises that the opportunities presented by GenAI and technology more broadly in the Australian education system are exciting. Some sectors have progressed well with integrating GenAI solutions, and the education sector as a whole needs to catch up to ensure that individuals and Australia can reap the technology’s potential benefits. It is imperative that we do not waste this opportunity and fall behind other countries.
4.97Stakeholders disagreed however about whether certain impacts of GenAI in education give rise to risks or opportunities. For example, some argue that there are risks that GenAI could increase teacher workload, while others contend that it will reduce workload and improve productivity. Policy making needs to account for this uncertainly—especially given the rapid uptake of GenAI and limited data on the longer-term effects—and balance potential risks and benefits.
4.98The Committee encourages the Australian Government to work with schools, TAFEs, and universities to encourage educators and students to employ GenAI tools in the different ways highlighted by UNESCO. For example, as a possibility engine, collaboration coach, and motivator.
4.99Educational institutions and users need assistance in understanding how to take advantage of what GenAI can offer in a safe and regulated environment. While more evidence and safeguards are required, some prospects include using data-driven insights to assist in educational contexts, creating administrative efficiencies, and reducing educators’ workload.
4.100The Committee sees the value of personalised learning to enable students to access resources and build knowledge and skills, which are tailored to their needs and progress. Individuals can access diverse resources and knowledge, tailored to their specific needs and development. However, given the possibility that students could become over-reliant on the technology—not only for study and assessments, but for their daily lives—it is the Committee’s view that GenAI must be used as a single resource and that the centrality of the human educator remains. Teachers, not technology, must continue to be the primary educator.
4.101The Committee heard about the extensive issue of equity of access to GenAI tools. The Australian Government needs to ensure that students in schools, TAFEs, and universities have equitable opportunities to understand and use GenAI tools ethically, safely, and responsibly. This is important based on principles of fairness and an equal go, to level the playing field, and to build an inclusive future for all Australians, acknowledging it will potentially impact career prospects and quality of life. Equitable access to GenAI tools is also about helping to future-proof Australia.
4.102Equity and access issues include having the infrastructure and hardware to enable the use of GenAI. The Committee recognises that for students and educators to use GenAI, they need access to a reliable internet connection, and to hardware like laptops or mobile phones, and that this access can be a challenge for marginalised communities.
4.103Equity is also about how GenAI is evenly integrated into schools, TAFEs, and universities, and the training that comes with it, so that it can be used appropriately and beneficially. In this transition phase, the Committee heard that educators may look to, or rely upon, the practices of earlier adopters in educational institutions. There is a need to ensure that all staff and students have help in accessing and navigating the technology. There cannot be a drop in standards nor a lack of support for cohorts already experiencing disadvantage.
4.104Stakeholders recognised the role of government and other key players to actively step in to overcome barriers to access to GenAI. The Committee heard that the digital divide could be exacerbated by GenAI—for instance, certain cohorts not having access to high-quality tools or training—or lessened by the technology—for example, by having an AI study buddy to help in areas with teacher shortages for subjects like STEM and languages. A national roll out of a GenAI tool that is high‑quality—not a lower baseline product—means disadvantaged schools can be on a more even playing field to better-resourced counterparts.
4.105The Committee is encouraged by the potential opportunity that GenAI presents to benefit the educational experience and outcomes of all students, including by creating additional value students with a disability or learning difficulty; students from low socio-economic or ESL backgrounds; Aboriginal or Torres Strait Islander students; students living regionally or remotely; and mature-aged students if they are time poor. GenAI can offer specific benefits to each of these diverse cohorts, from translation, to personalised learning plans tailored to a particular learning challenge, to access to education regardless of place or time. It is the Committee’s view that this these benefits to diverse cohorts should be actively supported.
4.106The Committee recommends that the Australian Government invest in training to teach educators of marginalised student cohorts—including disability or learning difficulty, low socio-economic, ESL, and regional, rural or remote—about how GenAI can specifically aid them.
4.107The Committee recommends that the Australian Government, in conjunction with educational providers, encourage educators and other staff to use GenAI tools for appropriate tasks to help streamline parts of teaching and administration.
4.108For example, lesson planning, timetabling, reporting, and simple grading for yes/no or multiple-choice questions.
4.109The Committee recommends that the Australian Government, in conjunction with educational providers and educators, use data-driven insights from GenAI tools for beneficial purposes.
4.110This includes to provide individually tailored feedback for students, to respond to identified systemic trends, and to streamline assessment processes.