Optimal engagement with AI in higher education teaching and research training

Abstract AI regulation is under active discussion, yet there have been no decisions about policy frameworks or legislation. Australia’s higher education (HE) regulatory model balances accountability with innovative capability. This chapter explores critical theory relating to AI, and analyses how leaders and policymakers need to clearly position AI to provide optimal engagement with AI in higher education teaching and research training. Five case studies of HE institutions in Australia are presented to contexualise the spectrum of institutional responses to the use of AI. The chapter draws on personal reflections of the author as a senior academic involved in course development and post graduate research supervision at several Australian HE providers. Document analysis of publicly available secondary data and literature triangulates the data. While researchers currently facilitate teaching and learning scholarship, research and publication using various mandated AI, there is little to minimise the risks of a breach of research integrity and undermining of learning within HE. By fostering collaboration among stakeholders, ensuring ethical alignment, and implementing ongoing monitoring mechanisms, institutions can navigate the intricate landscape of AI integration while enhancing educational opportunities for students. An effective response to upholding research integrity while using AI in HE requires sector-wide governance, flexible regulation, and collective advocacy.

Key Words Artificial Intelligence, research integrity, critical theory and higher education,

1 Introduction

Australia’s higher education (HE) regulatory model, administered by the Tertiary Education Quality and Standards Agency (TEQSA) balances accountability with innovation capability. Artificial Intelligence (AI) regulation is under active discussion and implementation across the Australian higher education (HE) sector, however, there have been no definite decisions by the government or industries about AI in education policy frameworks or legislation. An effective response to the of AI in research, teaching and learning in HE requires sector-wide governance, flexible regulation, and collective advocacy.

This chapter explores critical theory relating to AI in education, and analyses how leaders and policymakers in HE currently positions AI to suit their needs rather than providing optimal engagement with AI in higher education teaching and research training. Critical theory (Benson et al., 2017; Malott & Portfilio, 2011) views AI not merely as a technical tool, but as a socio-political phenomenon shaped by and shaping societal power structures. This chapter explores critical theory relating to AI in education, and analyses how leaders and policymakers in HE currently positions AI to suit their needs rather than providing optimal engagement with AI in higher education teaching and research training.

Five case studies are developed using organisational document analysis to explore the managerial and leadership dynamics in the creation of frameworks to manage AI in HE. The analysis draws on a decade of recent experience in course development and Higher Degree Research (HDR) supervision and mentoring in social work and community services with three universities and four international private providers in Australia. Academics facilitate teaching and learning scholarship, research and publications using various mandated AI, yet from the case studies provided here there is very little to minimise the risks of research misconduct and undermining of learning within HE in Australia and internationally.

The situation with AI in education is particularly dire for learners and researchers from historically disadvantaged cohorts because of their lack of access to up-to-date ICT equipment and software and to information and mentoring around use of AI and software in learning and research processes. An evidence-based framework is proposed that balances the opportunities of AI in research and learning processes while mitigating the risks to individuals and Higher Education Providers (HEPs).

2 The Regulation of AI in Australian Higher Education

Regulators have been late to the education game in the provision of policies and technical input into the use of Artificial Intelligence (AI) in higher education teaching and learning; the AI horse has bolted. Australia has become one of the most active global hubs for AI education and research, with universities investing heavily in teaching, research institutes, and national policy leadership.

In June 2024, TEQSA issued a Request for Information (RFI) to all Australian higher education providers requesting that they establish credible institutional action plans addressing the risk generative artificial intelligence poses to education and research integrity. TEQSA received a 100% response rate to this request. They issued two toolkits for HEPs – Gen AI strategies for Australian higher education: Emerging practice (TEQSA, 2024) and Gen AI Strategies for research training: Emerging practice (TEQSA, 2025a). The research toolkit aimed to provide guidance on Induction, guidance and training research processes, assessment and thesis examination and publications and grant applications.

These guidelines allow HEPs to then report on an annual basis on their progress in developing institutional actions plans. There are several relevant standards within the Higher Education Standards Framework (Threshold Standards) that providers may wish to consider in developing their institutional action, however, it is unclear what the ramifications are if HEPs do not have a “credible institutional action plan” or have an action plan that is deemed “not credible”. This whole emerging process of requesting HEPs to describe and self-regulate the use of AI and Gen AI is very opaque.

The TEQSA AI resource hub (TEQSA, 2025b) is the generative AI section of the regulator's higher education good practice hub, and it holds five kinds of material: the 2023 assessment reform principles, the 2025 guidance on enacting those principles, the June 2024 request for information and its follow-up, sector case studies and examples, and alerts about emerging risks such as contract cheating services built on AI. None of it is a Threshold Standard. All of it tells a provider how TEQSA reads Standards 1.4, 5.2 and 6.3 when generative AI is in the room.

A structured, in-depth overview of what AI and research education looks like across Australian universities in 2026 reveals that Australian universities are advancing AI on two fronts. Research into AI itself (machine learning, computer vision, intelligent systems, etc.) and using AI as a research tool across all disciplines. Australia produces world-class AI research, with 22% of its AI publications ranking in the top 10% globally, outperforming China and matching the US (Universities Australia, n.d.). Universities are also developing guidelines for ethical AI use, supported by TEQSA, ARC, and NHMRC.

AI Research Institutes such as the Australian Artificial Intelligence Institute (AAII) at the University of Technology Sydney (UTS) is one of the most prominent AI research centres in the country. AAII focuses on researching the theoretical foundations of AI, machine learning and computational intelligence, brain–computer interfaces, biomedical data science and intelligent drones and large-scale network analytics. It also runs research labs, seminar series, and major ARC-funded projects in autonomous learning.

The top Universities for AI Research in Australia according to EduRank (EduRank, n.d.) are the University of Melbourne, University of Sydney, University of New South Wales (UNSW), University of Queensland, Australian National University (ANU), Monash University and University of Technology Sydney (UTS). However, according to U.S. News (Global AI Ranking) those most prominent in “applied AI” are University of Technology Sydney, University of Sydney, Adelaide University, Monash University, Australian National University, UNSW, Deakin University, University of Queensland and RMIT University. These institutions are recognized for both research output and global reputation in AI.

AI as a topic for research training at Australian universities are actively around training HDR (Higher Degree by Research) students in AI methods; developing ethical AI guidelines, facilitation of interdisciplinary AI research groups and partnering with industry (e.g., Nvidia, Microsoft, Meta, Apple) to support innovation. The Australian National University (ANU) in Canberra has established a national working group to assist all universities to adapt to GAI and develop consistent policies. The group brings together senior leaders and experts in the HE sectors to advise on the constructive and ethical use of AI and GAI. The Australian higher education sector is known for strong global rankings, high research output and international collaborations, as well as Federal and State Government investments in AI capability in partnership with a growing tech sector and AI-driven start-up ecosystem with a clear focus on ethical and responsible AI.

As AI transforms workplace practices across industries, work-integrated learning approaches must evolve to prepare students for AI enhanced professional environments. Institutions should collaborate with industry partners to identify emerging AI proficiencies required in different fields and develop appropriate learning experiences (for an overview, see Dean et al., (2025)). While significant attention has been paid to the challenges AI poses to research integrity, less focus has been given to how AI might strengthen research integrity systems. Future development should explore how AI can support educational approaches to research integrity, including early identification of students requiring additional support.

ASCE at Deakin University has developed a framework that identifies several areas requiring further development as AI technologies evolve (Lodge et al., 2025). AI governance and institutional policy - the rapid and often fragmented adoption of AI tools across institutions presents significant governance challenges. Future work should focus on establishing clear, transparent, and agile governance frameworks. These frameworks must address key principles such as accountability, data privacy, ethical procurement, and the responsible use of AI by both staff and students, ensuring a cohesive and principled institutional approach.

As AI becomes increasingly embedded in professional practice across disciplines, ethics education must evolve to address emerging challenges. Future development should focus on how discipline-specific ethical frameworks can be adapted to address AI-specific concerns. The intersectionality of professional ethics, research ethics, personal ethics, and morals highlights the complexity of ethical decision-making. By acknowledging how these areas overlap and influence each other, individuals can strive for a holistic and integrative approach to ethical practices that supports justice, compassion, and respect for all stakeholders involved. This intersectional perspective enables a more nuanced understanding of ethical dilemmas and promotes a commitment to ethical integrity in diverse professional and research settings.

Intersectionality is a framework that explores how various social identities (such as race, gender, class, and sexuality) overlap and interact, influencing individuals' experiences and perspectives. This concept can be applied to various ethical domains, including professional ethics, research ethics, and personal ethics. Each of these areas has distinct principles and considerations, yet they are interrelated in shaping moral reasoning and behaviour.

3 Critical Theoretical Approaches to AI in Education

The current discourse(s) of AI in education focuses on the operational aspects, while the learning fundamentals are largely overlooked. Risks and opportunities around the use of AI in research and education manifest differently across various learner profiles, from K-12 students to professionals, with implications extending beyond individual learning outcomes to organizational effectiveness and information quality in broader societal contexts. A theoretically grounded perspective that can inform more nuanced policy approaches to balance technological advancement with educational effectiveness across diverse global contexts.

A critical theory lens (Waelen & Lindgren, 2025) focuses on locating and analysing the power dynamics, social injustices, and ideologies that shape societal structures. The link between education and technology is now both intrinsic and curious. When applied to the use AI in research in higher education, critical theory examines how AI technologies influence teaching practices, research and learning environments, and thus the broader social implications, the economic and social capital impacts. Critical theory aims to promote human emancipation and social justice by identifying and addressing factors that perpetuate inequality. In the context of AI, it encourages educators and researchers to critically examine how AI applications might reinforce existing biases or create new forms of inequality. For instance, biases in data used to train AI models can lead to unequal outcomes for marginalized groups, thus reproducing social inequalities within educational settings (Benson et al., 2017; Blair & Medina, 2016; Crompton & Burke, 2024a).

Analysing how power is distributed in educational environments creates understanding of the impact of AI on both educators and learners. Critical theorists argue that the adoption of AI in education must not strip away the agency of educators and students. Instead, AI should be employed to enhance human capabilities and support collaborative learning rather than serving purely as a tool of control and surveillance (Tobolowsky, 2014; Blair & Medina, 2016; Benson et al., 2017; Bailey, et al., 2010; Crompton & Burke, 2024a). Artificial intelligence is transforming adult learning by enabling personalized, adaptive, and accessible educational experiences that enhance skill acquisition and lifelong learning. AI allows adult learners to receive tailored educational experiences that adapt to their individual needs, learning pace, and prior knowledge. AI systems create "learning fingerprints" for each user, improving engagement, knowledge retention, and course completion rates compared to traditional online courses. Adaptive learning platforms analyse performance in real-time, adjusting content and suggesting resources based on strengths and weaknesses, which is particularly effective for adults balancing work, family, and education.

A critique of technological determinism challenges the notion that technology, including AI, is inherently neutral and that its outcomes are determined solely by the technology itself. Instead, it posits that the design, implementation, and usage of AI reflect the values and interests of those who create and control the technology. Critical theorists advocate for a reflective approach to technology integration, questioning the motivations behind AI applications and their alignment with democratic and ethical educational practices (Sanjakdar & Apple, 2024; Lowe, 2025; Nelson, 2024). The integration of AI in higher education is a complex process that requires careful consideration of critical theories and concepts to ensure AI is utilised effectively and within ethical guidelines.

3.1 The Role of AI in Facilitating Critical Thinking

Adopting a critical theory (Benson et al., 2017; Malott & Portfilio, 2011). lens when exploring AI in education allows for a more comprehensive understanding of the technologies' potential impacts on social equity, agency, and ethics. It encourages ongoing dialogue among educators, students, and researchers to critically engage with AI tools in a manner that contributes to a just and inclusive educational framework. By interrogating the design and application of AI, stakeholders can work towards harnessing its benefits while mitigating its risks and challenges.

Artificial Intelligence (AI) is increasingly recognized for its potential to foster critical thinking and learner engagement in educational settings. This integration leverages adaptive learning technologies and personalized feedback mechanisms to enhance students' analytical skills and involvement in their learning processes. AI can facilitate the development of critical thinking skills through various approaches and with the assessment of AI outputs, students can be encouraged to critically evaluate and analyse AI-generated content rather than merely using it as a means to an end. After generating ideas or solutions with AI tools, students might compare these outputs against their own reasoning, facilitating debates that foster deeper understanding and critical evaluation of various perspectives (Benson et al., 2017; Malott & Portfilio, 2011).

AI can support students in developing research skills by recommending diverse sources, however, educators should emphasize critical evaluation of these sources, promoting analytical skills as students’ fact-check and assess credibility beyond AI suggestions (Benson et al., 2017). “What If” Scenarios where educators can pose open-ended questions that require students to brainstorm without relying solely on AI. This method nurtures divergent thinking and encourages students to explore complex problems creatively (Benson et al., 2017; Totten & Pedersen., 2014).

AI-driven educational tools create opportunities for enhanced learner engagement via adaptive Learning Technologies: By analysing students' learning patterns, AI systems can deliver personalized learning experiences suited to individual needs. This adaptability helps maintain engagement by providing content that is neither too challenging nor too simplistic, thus optimizing students' learning experiences (Lovern & Knowles, 2015; Totten & Pedersen, 2014; Capper & Henry, 2025).

AI enhances learner engagement by providing real-time feedback on tasks and assessments. This timely input enables students to recognize errors promptly and encourages active participation in correction, reinforcing concepts as they learn (Totten & Pedersen, 2014; Elkhodr & Gide, 2025; Bailey et al., 2010). AI tools can incorporate gamified elements into learning. Games designed to improve problem-solving skills not only make learning enjoyable but also promote sustained engagement by encouraging competition and collaboration among peers (Benson et al., 2017; Babu & Naithani, 2024; Totten & Pedersen, 2014; Elkhodr & Gide, 2025; Bailey et al., 2010).

While AI offers notable benefits, there are challenges to consider the overreliance on AI. Excessive reliance on AI for information retrieval and problem-solving could lead to passive learning habits, where students do not critically engage with content or develop their own reasoning skills. Educators must find a balance that encourages students to be active learners rather than passive consumers of information (Totten & Pedersen, 2014; Babu & Naithani, 2024; Lovern & Knowles, 2015).

There is a risk that AI systems will perpetuate human biases if not properly designed. This can affect the critical assessment of AI-generated outputs, highlighting the need for students to develop skills to identify and interrogate biases in data and recommendations provided by AI (Capper & Henry, 2025; Malott & Portfilio,2011; Nelson, 2024). Privacy concerns around the use of AI involve issues such as significant data collection, raising questions about data privacy and ethical use. Institutions have a responsibility to ensure that AI technologies are implemented in ways that safeguard students' data and promote inclusion and equity (Sanjakdar & Apple, 2024; Totten & Pedersen, 2014).

AI does have the potential to both facilitate critical thinking and enhance learner engagement through personalized learning experiences, immediate feedback, and invigorating educational environments. However, for its full benefits to be realized, careful consideration must be given to potential challenges such as overreliance on technology and ethical concerns regarding data use. As educators integrate AI into their practices, balancing the human elements of teaching with the capabilities of AI will be crucial for fostering critical thinking and engagement in learners.

In relation to curriculum design and pedagogy educators need to critically analyse how AI can reshape curriculum and pedagogical practices as andragogy or ‘polygogies’ (Veever- Williams, 2025). A critical theory lens emphasizes the importance of engaging students in discussions about AI, its implications, and ethical considerations. This involves fostering critical thinking skills and encouraging students to question the biases inherent in AI systems (Chigona et al., 2024; Crompton & Burke, 2024).

Ethical and responsible use of AI in education must be prioritized. This includes ensuring transparency in AI systems, protecting student data privacy, and implementing practices that prevent algorithmic biases. Instructors and institutions should foster an ethical culture around AI usage, guiding students in understanding their rights and the societal implications of AI technologies (Elkhodr & Gide, 2025; Guarcello & Longo, 2024; Sanjakdar & Apple, 2024; Totten & Pedersen, 2014).

Integrating generative AI into higher education is crucial for the preparation of future generations of GAI-literate students and workers and entrepreneurs. A comprehensive analysis, globally, of higher education institutional adoption of policies managing the use of AI remains absent, specifically any with a theoretical lens. Australian universities are proactively addressing AI integration by emphasising research integrity, enhancing teaching and learning practices, and promoting equity.

Addressing the digital divide is crucial and critical theory lens directs us toward equitable access to AI tools, ensuring that all students benefit from personalized learning experiences while also learning to navigate the challenges that these technologies can introduce. Policies and practices must be implemented to prevent AI from widening existing educational disparities (Corbeil & Corbeil, 2025; Dervin & R’boul, 2025; Oruganti, et al., 2025).

4 Australian Higher Education Case Studies

The following five (5) institutional case studies, the fifth containing two (2) Higher Degree Research (HDR) vignettes, derive from the author’s own experiences of working with current and new higher education providers (HEP) in Australia as they develop new courses in the era of AI and as they expand their HDR degrees in community services and social work. The information is current (2022-2026) and informs the author’s own reflections and practices in the use of AI in both creating learning frameworks and participating in HDR and community- engaged research.

4.1 Methodology

Document analysis of publicy available organizational data was used in this research in combination with participant observer reflections as a means of triangulation of data to confirm the phenomena of policies and procedures within the international education context of use of AI (Pelizzon et al., 2023). Document analysis (Sarantakos, 2013) is a systematic process for evaluating data from documents, including printed and electronic (computer-based and Internet-transmitted) material. Document analysis is data examined and interpreted to elicit meaning, gain understanding, and develop empirical knowledge about a specific topic or process. Documents contain text (words), numbers and images that have been recorded with or without a researcher’s intervention, data that can be referred to as ‘social facts’, which are produced, shared, and used in socially organised ways (Bowen, 2009).

The author’s experiences since 2015 have been in various roles, as a Head of School, course development consultant, as an academic dean, as a member of academic boards and as an accreditation expert in social work and community services and supervisor of PhD degrees. These experiences across more than a dozen higher education institutions suggested that higher education institutions, whether they were public universities or private, provided the most minimal policy and procedure frameworks possible around use of AI in research in teaching and learning. This chapter addresses the gap in HE policy and practice about the ethical development and use of AI to uphold research integrity.

Participant observation (Ingram, 2026) is a research method of data collection in which researchers participate in and observe events and actions within a specific research field. Participant observation involves the observer being a member of the setting in which they are collecting data. Participant observation as ethnographic research seeks to determine - what is going on here? It seeks to better understand and capture the context within which people interact. Firsthand experience with a setting allows researchers to open to discovery and inductive, rather than guessing about the context (Angrosino, 2005). The researcher usually discovers organisational mechanisms that routinely escape awareness of the participant using a different method, participant observation provides a chance to learn things that people may be unwilling to discuss in interview or survey type investigations (Angrosino, 2005).

Document analysis (Sarantakos, 2013) is a systematic process for evaluating data from documents, including printed and electronic (computer-based and Internet-transmitted) material. Document analysis is data examined and interpreted to elicit meaning, gain understanding, and develop empirical knowledge about a specific topic or process (Sarantakos, 2013). Documents contain text (words), numbers and images that have been recorded with or without a researcher’s intervention, data that can be referred to as ‘social facts’, which are produced, shared, and used in socially organised ways (Bowen, 2009).

The data for this chapter were a collection of curriculum and educational documents, archival materials and artifacts at each of the HEPs relevant to the research including, websites, curriculum documents, government standards, accreditation guidelines, meeting minutes, policies and procedure documents, emails, social media, reports, organizational research publications, and in-house blogs.

Observations of curriculum development documented included reflections of the researchers from meetings, emails, course development processes, course advisory meetings, organizational curriculum processes and reviews, TEQSA processes and reviews, and Australian Association of Social Workers (AASW) professional social work education accreditation processes.The author’s experiences since 2015 have been in various roles, as a Head of School, course development consultant, as an academic dean, as a member of academic boards and as an accreditation expert in social work and community services and supervisor of PhD degrees. These experiences across more than a dozen higher education institutions revealed that each higher education institution, whether they were public universities or private providers, were motivated to provide the most minimal policy and procedure frameworks possible around use of AI in research and in teaching and learning because of a reluctance to be interventionist. Also, there is an unwillingness to intervene in how academic researchers, educators and students utilised AI to managing learning and research.

Ethical issues for critical ethnographic research are like other forms of qualitative research: issues of informed consent, guarantees of anonymity, giving back to the organisation, are all issues to be addressed when analysing organisational issues. Additional issues such as committing ourselves as researchers to long-term relationships where trust is an especially important issue as we observe and hear things of a sensitive nature, and we may become a sounding board’ for the participants. Human Research Ethics approval was not required as no individuals were interviewed or had their own data utilised in this process. This research analyses the personal and professional reflections of the author, his experiences of AI in higher education and publicly available secondary data. The case study institutions have been de- identified in this chapter to preserve anonymity and their documents have not been cited in this chapter.

4.2 Case studies

Institute #1 is an established private provider who offers courses in allied health. In 2022, the institute had no policy or procedural frameworks for regulating or monitoring AI use by staff or students. Tensions remain in 2026 for this institution around managing research integrity versus progression of students through course to produce graduates who are work ready but not necessarily academically or ethically proficient is use of AI. This HEP is currently undergoing course accreditation review by TEQSA to audit and scrutinise their management of teaching and learning processes.

Institute #2 is a new higher education provider This institute as part of its application for registration as a new HE provider developed a minimalist framework reading AI which stated that assessments are designed to enable students to responsibly use AI and other enabling tools in completing assessment activities while abiding by the Institute’s policies and procedures, e.g., Assessment Policy and Procedure and the Student Academic Integrity Policy and Procedure.

It is clearly stated in all the documentation that students’ assessments and work must be their own and where the use of AI tools, such as ChatGPT, have been permitted by the course / subject lead, the use of software and other AI resources must be properly credited and referenced, and all assessments submissions will need to be seen as substantially students’ own work. AI is seen by this institute as a tool to source information and a legitimate source which must be referenced alongside all other sources of information and data. Examples cited in academic library guides include:

Referencing Format: Author or company of AI tool. (Year, Month Day). Title of the chat. [Description of the chat type]. AI tool or model. Web address of the chat.

Example: OpenAI. (2026, January 10). Methods of providing alterative text for accessibility on websites. [Generative AI chat]. ChatGPT-5. https://chatgpt.com/share/98b8790-1tt3-400g- 3nhh-it6cs923t977

Students and staff must then substantially demonstrate that in their assessments, their research proposal, or their theses that the information gained from AI has been adapted to fit the work being produced, just like any other source of material. This is a simple framework or managing use of AI with the onus of educators, teachers, research supervisors and academic deans to manage the processes of academic integrity.

Recent changes in the assessment framework are designed to reduce surface-level assessment load, increase meaningful in-class engagement, support the faculty's ability to assess student readiness for fieldwork placements, meet international student visa attendance obligations, and better align the program's assessment practices with the holistic learning outcomes appropriate for community services graduates.

Institute #3 is an established provider which offers courses in management, IT and human services and has established a clear understanding at management level of AI leveraged as advanced machine learning models to gather available data (both valid and otherwise) and generate some form of content, appearing to simulate human intelligence. Using AI tools (or tools that have AI built into them) are seen as new opportunities in processes of how to learn and conduct research. Use of AI is stated to staff and students as being tools that can assist everyone with time management, organisation and academic skills, however, when used inappropriately, these same tools can compromise the integrity of the work and undermine learning.

It might be challenging to decide how and when to use AI efficiently and ethically. To this end, an on online module was made available for all staff and students to complete which has been designed to provide both staff and students with all the guidance on what constitutes appropriate and inappropriate use of AI tools in learning and assessment tasks at the Institute. Essentially, copying of any materials from anywhere was considered a breach of the Institute’s academic integrity policies and procedures and the penalties laid therein.

At Institute #3, students and researchers are not permitted to use AI tools to generate academic content, unless explicitly required as part of an assessment task. In this context, “to generate academic content” means using these tools to develop an assessment, or part of an assessment. Staff and students were instructed to be aware that using any AI tool to generate academic content without explicit authorisation from each lecturer or subject coordinator would represent a breach of academic integrity and equate to academic misconduct, as explained in the relevant policies and procedures and as outlined in the online module.

All use of AI tools is thus contextual and subject dependent. This practical, flexible approach to AI uses reflects the demands of each subject: some uses of AI could prevent you from meeting the learning outcomes for a specific subject, but not for another. Often, the limits on AI use are stated in the learning and assessment guides. This will vary across subjects and across assessments within the same subject, so staff and students are advised not make any assumptions.

Institute #4 is a new provider, and. is currently addressing the use of AI by providing a comprehensive suite of policies, procedures and a formal AI Plan for adoption because it sees AI as a global disruptor, in a positive sense. The Australian Higher Education Standards Framework from 2025 has required HEPs to address the use and management of AI via a draft planning document, to ensure students have current workforce readiness, and to facilitate academic and student expectations about the use of AI in their studies.

The AI Plan aims to promote the guiding values of integrity of use of AI as research tools and for promoting learning but not for writing and analysis, innovation in the guiding and collaboration between academics, educators, professional staff and students in the use of AI in higher education. Equity in the use of AI is addressed whereby everyone has access to the tools that are allowable and useful for their courses, and inclusion, AI can create opportunities for people not currently accessing higher education studies due to neurodiversity’s or physical disabilities.

A culture of responsible use of AI by educators, staff, and students, requires cross-disciplinary collaboration and the effective integration of AI requires collaboration between academics, staff, administrators, and IT professionals. This collaborative approach ensures that AI development and implementation align with the institution's educational goals and ethical principles.

Transparency regarding the use of AI, including their limitations and potential biases, is crucial for building trust within the academic community. Transparency is essential for the education community, including leadership, educators, staff, and students because it underpins the ethical use of AI, fostering an environment of trust, accountability, and open dialogue with all constituents.

Ongoing training and support for education staff is also essential as AI continues to evolve and as such are essential to ensure that all educators can maintain their AI literacy, that is, the ethical, pedagogical, technical dimensions of AI and how to adapt to new advancements. This ensures that all members, including academic leaders, stay current with AI advancements and can lead effectively in adapting these technologies to all educational settings. Such ongoing education supports the Institute’s commitment to academic excellence and leadership in innovation.

Institute #5 is a university with a Responsible AI Adoption Policy which outlines that the University’s approach to AI demonstrates their organisation cultural qualities. Being accountable by applying clear lines of responsibility and instil transparency in AI use. Being connected and engage across disciplines and functions, using cross-University collaboration to identify, manage and monitor issues and risks. Innovation by applying a risk-based approach, classifying AI according to its level of risk and applying corresponding levels of oversight and controls to mitigate that risk. Providing care to actively consider environmental sustainability and the impact of unwanted bias, data security and potential benefits in the use of AI. Although they may be supported by AI applications, significant decisions will be made by humans to promote a safe and inclusive community. The university’s policy document is comprehensive and prescriptive and provides clear guidelines for all faculty staff and students on risk minimisation of use of AI.

These five (5) case studies demonstrate that in 2026 there is widely variant responses to the use of AI in teaching and learning and research in higher education in Australia. There is no comprehensive single blueprint for use of AI but there is a reliance by regulators and governments on higher education institutions governing their own processes and outcomes.

5 Critical Theory Analysis of AI Policy Development

A critical theory analysis would examine not only whether each institute’s AI rules are clear or effective, but also whose interests they serve, how they distribute power and risk, and what they treat as legitimate knowledge and learning. Drawing on the Frankfurt School, Habermas, Foucault, Freire, and critical scholarship on technology and equity (Delanty & Harris, 2021), the cases reveal a tension between education as human development and education as a regulated, market-facing system expected to deliver compliant, “work-ready” graduates. The cross-cutting themes are as follows.

Instrumental reason versus education as a public good. Across the cases, AI is often framed to create efficiency, compliance, employability, or institutional competitiveness. A critical theory perspective asks whether these goals displace deeper purposes of higher education: developing judgment, independent inquiry, ethical reasoning, and the capacity to question existing arrangements. AI can support learning, but when its value is measured mainly by speed, progression, or workforce readiness, education risks becoming instrumental focused on producing outputs rather than cultivating understanding.

Power and governance. AI policies do not merely regulate tools; they define acceptable conduct and make some forms of knowledge production visible and punishable. Institute #1’s absence of policy is not neutral: it leaves decisions to informal practice and can make standards inconsistent. By contrast, Institutes #2–4 formalise rules, but those rules also concentrate interpretive power in educators, managers, and compliance systems. A key question is whether students and staff have a meaningful role in shaping the rules that govern them.

The politics of authorship and knowledge. The cases assume that work can be divided into “the student’s own” and material derived from AI or other sources. Critical analysis complicates that distinction: knowledge is always produced through prior sources, tools, collaboration, and institutional conventions. The important questions are what learning or intellectual contribution an assessment is intended to demonstrate, how that contribution can be evidenced, and whether rules about attribution are consistently applied to human and machine sources.

Surveillance, trust, and due process. AI detection scores and other software can turn assessment into a process of monitoring and suspicion. A similarity or “likeness” score is not, by itself, proof of misconduct. Fair governance requires transparent procedures, human review, an opportunity for students to respond, and attention to the limitations and possible biases of detection systems. Otherwise, technical tools risk substituting for educational judgment. An Institute-by-institute analysis provide more detail.

Institute #1’s policy absence and normalised permissiveness illustrates institutional omission as a form of governance. Without a framework, staff and students may develop practices by default, while the institution avoids making explicit decisions about authorship, disclosure, assessment validity, and responsibility. The resulting permissiveness is not necessarily student-centred: students may be passed without clear evidence that they have met learning outcomes, while responsibility for interpreting acceptable AI use is effectively displaced onto individuals.

Institute #2 is attribution-centred, but potentially too procedural but offers a clearer model: AI may be used where permitted, must be acknowledged, and submitted work must remain substantially the student’s own. This supports transparency and gives students a stated framework. Yet a critical theory lens asks whether citation is being treated as a sufficient proxy for learning and integrity.

Referencing an AI output can disclose its use, but it does not establish that a student evaluated it, understood it, or can apply the underlying knowledge. Nor does it resolve whether a particular task is meant to assess the student’s unaided reasoning, practical skill, or ability to use AI critically. “Substantially their own” also requires careful definition; otherwise, the phrase may leave significant discretion with assessors and create uneven enforcement.

The framework places much of the burden on educators, supervisors, and deans. That may enable contextual judgment, but it can also produce inconsistent expectations across subjects. A more dialogic approach would involve students and staff in defining permitted uses and would explain the learning purpose behind each restriction or permission—not only how to cite the tool.

Institute #3 has contextual flexibility and the risk of discipline and recognises that AI’s appropriateness depends on the subject and learning outcomes. This is pedagogically stronger than a single blanket rule: an AI use that undermines one assessment may be an appropriate object of study in another. Its staff-and-student module also signals a commitment to shared understanding.

However, the default prohibition on using AI to generate academic content, unless explicitly authorised, can create a permission-based regime in which innovation depends on individual lecturers or coordinators. If expectations vary between subjects or assessments, students may face uncertainty despite being told not to make assumptions. The policy may also privilege students who are already familiar with academic conventions and confident in asking for clarification.

From a Foucauldian perspective (Lioliou & Willcocks, 2019), the misconduct rules can be understood as disciplinary mechanisms: they teach appropriate conduct while making students’ work subject to institutional classification and sanction. Such rules may be necessary, but their legitimacy depends on clarity, proportionality, consistent application, and opportunities for students to understand and contest decisions. The institution should also distinguish between prohibited substitution for learning and legitimate uses such as accessibility, planning, language support, or critique of AI outputs.

Institute #4 has an ambitious inclusion agenda and implementation tensions with a plan that is the most comprehensive and explicitly connects AI governance to accessibility, equity, transparency, staff development, data security, and institutional capacity. This is a significant strength. Its recognition that unequal access to tools and training can reproduce disadvantage moves beyond treating AI as an individual student’s integrity issue.

At the same time, the plan contains tensions that require resolution. It promotes AI as a positive disruptor and supports workforce readiness, but its stated integrity values limit AI’s role in writing and analysis. That raises practical questions: which activities count as “research tools,” and where does information gathering end and analysis begin? The plan’s commitments to accessibility and inclusion may also conflict with restrictions if students rely on AI-enabled tools for communication, organisation, or assistive support. A blanket or poorly specified restriction could disadvantage disabled students, neurodivergent students, and multilingual learners.

There is also a possible gap between extensive planning and everyday practice. A detailed plan can become a managerial document rather than a shared educational commitment unless staff have time, resources, and authority to implement it—and students can influence how it operates.

6 Frameworks to guide responsible development and use of AI

Artificial Intelligence (AI) brings transformative potential across various sectors, yet its deployment raises significant ethical, regulatory, and societal challenges. Consequently, frameworks that guide responsible AI development and deployment are crucial. Taken together, the cases show that the central issue is not simply whether AI is allowed or prohibited. It is how institutions define learning, authorship, competence, fairness, and accountability when AI is part of knowledge production. Institute #1 shows the consequences of leaving those questions unresolved; Institutes #2 and #3 show different attempts to regulate practice through attribution and contextual permission; Institute’s #4 and #5 make the broadest effort to connect AI governance with institutional values. A critical theory-informed approach to AI would:

• design assessment around the capabilities students are expected to demonstrate, rather than rely primarily on AI detection;
• state permitted and prohibited uses in task-specific, accessible language;
• distinguish disclosure from proof of learning or competence;
• provide fair, transparent processes for investigating suspected misconduct;
• involve students, educators, professional staff, and relevant communities in policy development;
• protect legitimate accessibility and language-support uses while preventing AI from substituting for required learning;
• evaluate whether policies produce equitable outcomes, not merely whether they are documented.

The aim would not be to maximise AI adoption or eliminate it, but to ensure that AI use remains answerable to educational purposes, professional responsibilities, and the interests of learners—rather than allowing efficiency, surveillance, or compliance to define higher education on their own. Key policy measures include the development of guidelines for ethical AI use, the design of authentic assessments to mitigate misuse, and the provision of training programs for faculty and students to foster AI literacy. Gaps remain in comprehensive policy frameworks, particularly in addressing data privacy concerns and ensuring equitable access to AI tools. What is required are actionable insights for policymakers to craft inclusive, transparent, and adaptive strategies for integrating AI into higher education.

Ethical AI frameworks aim to ensure AI technologies are designed and used in ways that are fair, transparent, and aligned with societal values. Some influential frameworks include:

• EU Ethics Guidelines for Trustworthy AI: This framework emphasizes several key principles, including transparency, accountability, and privacy (European Commission, 2019). It highlights the importance of maintaining human oversight in AI systems to guard against unintended consequences (Totten & Pedersen, 2014; Tobolowsky, 2014; Bailey et al., 2010).
• IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems: This approach advocates for values-based design, promoting ethical considerations in AI development across multiple disciplines, thus ensuring interdisciplinary collaboration (Crompton & Burke, 2024; Brooks, 2024).
• Responsible AI in Healthcare: Specific to healthcare, this framework highlights the importance of safeguarding patient data and ensuring AI systems operate transparently and fairly. Upholding these ethical standards is paramount as the integration of AI in healthcare evolves (Crompton & Burke, 2024; Tripathi & Rosak- Szyrocka , 2024).

Despite the potential advantages, deploying AI brings considerable risks, challenges include bias and discrimination - AI systems can perpetuate existing biases if trained on flawed data. Careful attention to data selection and algorithm transparency is essential to mitigate these risks (Elkhodr & Gide, 2025). Privacy and data security when handling large datasets in AI applications raises significant concerns; ensuring robust data protection frameworks is important to secure sensitive information (Elkhodr & Gide, 2025). The lack of transparency means many AI systems operate as "black boxes," making it challenging to understand how decisions are made. Increasing transparency around AI processes is crucial for maintaining public trust and accountability (Lang-Raad, 2024).

7 The way forward for managing AI in HE

The integration of AI into various contexts necessitates a careful and respectful approach to its deployment. By employing ethical frameworks, acknowledging risks, and fostering stakeholder engagement, organizations can harness the benefits of AI while minimizing potential pitfalls. The message is clear: in the pursuit of innovation, it is essential to Stop, Pause, and Move Ahead with Caution, ensuring that AI promotes societal good without compromising ethical values.

As AI continues to infiltrate various sectors, its influence in higher education demands diligent management. To navigate the complexities of AI integration, institutions must consider comprehensive frameworks, advisory roles, and potential gatekeeping measures aimed at ethical and effective implementation.

7.1 Balancing Innovation and Caution

Higher education providers should adopt a “Stop, Pause, Move Ahead with Caution” philosophy when innovating with AI. Phases of development should include:

1. Pausing for Ethical Assessments = before deploying AI systems, organizations should conduct thorough ethical impact assessments to identify potential risks and ensure systems align with ethical guidelines (Oruganti et al., 2025; Lang-Raad, 2024).
2. Continuous Monitoring by establish robust monitoring mechanisms to assess AI system performance regularly. This includes checking for bias, transparency, and alignment with ethical standards throughout the AI lifecycle (Lang-Raad, 2024).
3. Involve diverse stakeholders in the AI development process, including ethicists, policymakers, and community representatives, to evaluate the societal impact of AI technologies and adapt frameworks accordingly (Oruganti et al., 2025; Murugan et al., 2025).

7.2 Engaging with regulatory frameworks and developing purposeful blueprints

The evolving legal landscape surrounding AI highlights the importance of compliance with national and international regulations. Data Protection Laws - Regulations like the GDPR in Europe focus on data protection in AI applications, offering a legal framework that demands transparency and accountability in AI processes (Crompton & Burke, 2024). National AI Strategies - countries worldwide are developing national strategies that address AI's implications for societal welfare. These strategies often prioritize ethical considerations and collaboration among multiple sectors to enforce standards effectively (Oruganti et al., 2025; El- Bassiouny, et al., 2025).

The development of robust frameworks is critical for guiding the ethical deployment of AI in higher education. These frameworks can define ethical guidelines such as the EU's Ethics Guidelines for Trustworthy AI advocate for transparency, accountability, and ethical considerations in AI applications. Institutions should adopt similar standards to ensure AI aligns with educational values and societal norms (Brooks, 2024).

Integrating Best Practices Frameworks can encompass best practices for data collection, algorithm transparency, and bias mitigation. This includes clear protocols for training AI on diverse and representative datasets to prevent the perpetuation of biases (Oruganti et al., 2025). Fostering Interdisciplinary Collaboration by engaging stakeholders from various disciplines— including educators, ethicists, and technologists—can enhance the relevance and applicability of AI frameworks, ensuring they address specific challenges faced by institutions (Elkhodr & Gide, 2025).

Blueprints for AI implementation in higher education should focus on structured, phased approaches such as assessing organizational readiness where institutions can carry out readiness assessments to evaluate their existing infrastructure and identify gaps that need addressing before introducing AI technologies (Corbeil & Corbeil, 2025). Starting with pilot programs allows institutions to test AI applications in controlled environments, gathering insights and feedback for refinement before a wider rollout. This approach can help mitigate risks associated with large-scale implementation (Crompton & Burke, 2024). Blueprints should recognize the diversity of educational contexts and suggest tailored implementations that cater to the specific needs of different academic departments or institutions, ensuring a customized approach to AI integration (Lang-Raad, 2024).

Establishing advisory roles within institutions is crucial for informed decision-making regarding AI technologies. Formulate advisory committees is something institutions should consider creating, comprising faculty, industry experts, and ethics scholars to oversee AI implementations, ensuring they align with educational goals and ethical standards (Oruganti et al., 2025). Continuous professional development comprising training programs focused on AI ethics, technology integration, and data literacy for faculty can prepare educators to adopt and effectively employ AI-tools, promoting a more informed academic environment (Murugan et al., 2025). Student involvement in the advisory process can provide valuable perspectives and insights, fostering a culture of inclusivity and responsiveness to emerging needs and concerns regarding AI use in education (El-Bassiouny, et al., 2025).

While implementing AI can enhance educational experiences, organizations must also consider gatekeeping measures to protect academic and research integrity and ensure ethical usage. Monitoring and evaluation is where institutions should implement systems for ongoing monitoring of AI usage, assessing impact on student outcomes and ensuring compliance with ethical guidelines. Regular audits can help address any identified issues proactively (Crompton & Burke, 2024).

Developing clear policies regarding AI usage in academic settings—including guidelines on data privacy, AI-generated content, and assessment integrity—will help mitigate concerns related to academic dishonesty and misuse of AI technologies (Crompton & Burke, 2024). Facilitating Peer Review mechanisms to review AI-generated work can maintain academic standards and integrity. Encouraging a culture of scrutiny for AI use in assessments will help safeguard against overreliance on technology for academic tasks (Murugan et al., 2025).

8 Conclusion

The effective management of AI in higher education requires a multifaceted approach that encompasses comprehensive frameworks, purposeful blueprints, advisory roles, and effective gatekeeping measures. By fostering collaboration among stakeholders, ensuring ethical alignment, and implementing ongoing monitoring mechanisms, institutions can navigate the intricate landscape of AI integration while enhancing educational opportunities for students. Adopting this cautious yet progressive strategy will be instrumental in harnessing the full potential of AI in the academic sphere.

Tools like ChatGPT can enhance efficiency and accessibility, their use in and as writing remains stigmatized due to traditional notions of authorship and intellectual rigor. We need to consider advocating for a more balanced perspective on the use of AI in education that values content quality over outdated ideals of “textual purity” (Suchikova & Tsybuliak, 2025). Any discussion or analysis of the use of AI in higher education is going to be a snapshot in time as there is rapid development in the types of AI and the uses of AI in teaching, learning and research. We need to acknowledge the evolving role of AI in academic research by addressing the challenges, opportunities, and ethical considerations surrounding its use.

9 Limitations

This chapter is based on the experiences and personal perspectives of the author as an academic and consultant working across the higher education sector in Australia and across Asia, and subjective bias may be present. However, the arguments presented are supported by the data analysed including publicly available documents, literature and grey literature. The AI development in education is occurring at an exponential rate, and this chapter is therefore limited in its ability to fully include ongoing changes in AI use in higher education research settings. The proposed way forward counteracts this evolving AI landscape by focusing on adaptive approaches.

Acknowledgement of the use of AI

In preparation of this manuscript, the author used the Large Language Model ‘Gemini’ developed by Google to format references according to the APA 7th referencing style. The accuracy of all references was checked by the authors who remain responsible for all citations in this chapter.

Declaration of Competing Interests

The author has no competing interests to report in the content of this manuscript.

References

About the author

Professor Rob Townsend, PhD, M.Ed, BSW

Rob is Managing Director of Naturally Gifted Research & Associates, which provides higher education and social work academic advisory services, social work, and community services curriculum development, AASW accreditation panels for higher education providers, and social and community services work accreditation consultancy. He is currently Chair of Social Work at Alliance Higher Education Institute and at Lincoln Institute of Higher Education; AASW Accreditation Council Member, ANZSWWER Member, International Academic Forum Member, The Learner Academic Research Community Member. His research interests include adult education, ethnographies of communities, autoethnographies, grounded theory, critical social theories, diversity in urban and regional communities, adult transitions through the life course, trauma-informed community practices, social and human capital development, animal, and arts-based therapies.