Artificial intelligence (AI) is a key driver of the digitalisation that is transforming the accounting profession and broader business environment, requiring educators to adopt new technologies to remain competitive. AI’s rapid adoption across industries has increased expectations of efficiency, decision making, data analysis capability and potential for innovation.

In their book, Artificial Intelligence: A Modern Approach, authors Stuart Russell and Peter Norvig categorise AI into three main types, each reflecting differing levels of capability and potential application. These main types are:

1. Artificial Narrow Intelligence (ANI)

2. Artificial General Intelligence (AGI); and

3. Artificial Superintelligence (ASI).

Understanding these distinctions is important in assessing how AI can drive organisational efficiency and effectiveness, particularly within an educational context.

Former Stanford University Professor Harold J. Leavitt’s 1965 organisational model—”Leavitt’s Diamond”—identifies four important variables to assess within organisations, including task, structure, technological and human variables. These variables, Leavitt argues, must be aligned when implementing change.

Introducing AI into educational environments necessitates more than the deployment of tools; it requires the redesign of teaching processes, the development of employee capabilities and a reconsideration of how students engage with knowledge.

Without addressing these elements, AI adoption risks remaining superficial and failing to enhance meaningful learning. AI plays its role, but it is not the full solution.

In an article published in the Harvard Business Review in 2021, OnCorps AI founder Bob Suh suggests the key for decision-makers seeking to optimise their work with AI, is to identify those problems best suited to AI resolution as opposed to those requiring a managerial mind.

Equally, accounting lecturers must develop a sound understanding of AI capabilities while questioning the technology’s implications for teaching, learning and the development of competencies.

Opportunities in accounting education

Accounting lecturers need to design teaching and assessments in ways that make AI part of the learning process, rather than something students use separately or incorrectly. It should be a tool for learning—but should not substitute critical thinking.

As part of an assignment, our students were asked to utilise AI tools to support their work and subsequently share their experiences through questionnaires and discussions. The feedback offers valuable insights into how third-level accounting students are integrating AI into different stages of the assignment process.

Many report using AI across all stages of their assignments, including in research, drafting answers, analysing data, breaking down complex sustainability frameworks and refining their work. It is clear AI can be used very effectively as a support tool.

This is particularly evident in the analysis of sustainability reports, where students are required to engage with very long reports of over 400 pages in some cases. Here,

AI allows students to navigate these documents more efficiently and extract relevant information in a way that would otherwise be extremely time-consuming.

Students use tools such as Canva, NotebookLM and Gamma to improve the presentation of their work, producing more structured and visually appealing outputs.

In general, they feel that AI improves their efficiency and productivity, helping them better understand sustainability reporting concepts and supporting the overall quality of their work, particularly in terms of clarity and structure.

Many students report increased confidence in completing assignments, referring to AI as a “personal tutor”. This raises an important question, however.

While AI clearly supports students in completing tasks more efficiently, whether or not the technology is supporting genuine learning is less clear. There is a risk that students may rely on AI to interpret and summarise information, rather than developing these skills themselves.

As a lecturer, this leads me to question whether students are using AI as a learning tool, or simply as a way to produce work more quickly.

Students describe editing and deleting AI-generated content they deem repetitive or unsuitable. This suggests they are not passively accepting AI outputs in all aspects of their work but instead engaging with these outputs, at least to some extent.

The level of engagement may vary, however, and this selective use raises questions regarding the consistency with which AI outputs are critically evaluated.

Students express caution about sharing their personal information with AI tools, indicating an awareness of data privacy concerns. This is a positive development, suggesting some are beginning to consider the broader implications of AI use beyond academic performance.

Careful consideration must be given to how AI is used within teaching and assessment. If students are to benefit fully, AI must be used in a way that encourages engagement, critical thinking and understanding, rather than replacing these processes.

Assessments need to be designed in a way that ensures students are always asked to explain, justify and critically reflect on their work, rather than simply present it.

AI challenges in accounting education

In the context of accounting education, the use of AI is largely limited to Artificial Narrow Intelligence, with students relying on tools like ChatGPT and Microsoft Copilot to support tasks such as research, summarisation and presentation.

While these tools may enhance efficiency, their impact on effectiveness is less certain, particularly in terms of learning outcomes

One of the most immediate challenges I see with the use of AI in accounting education is that AI does not always produce the best or most accurate outcome. It is limited by the data it is given and can present information that is incomplete, misleading or simply incorrect.

This is evident in students’ work, where issues such as spelling errors and other inaccuracies arise. Although these mistakes may appear minor, they highlight a broader concern regarding the reliability of AI outputs and the potential for students to place too much trust in these outputs.

Occasionally, for example, AI will confuse fiscal years in reports and refer to financial figures that do not exist in the report. All work needs to be cross-referenced and verified.

While AI improves efficiency, it also raises a concern about how students develop research skills, particularly in terms of locating, evaluating and selecting appropriate sources.

The technology may also limit opportunities for students to develop their own writing and communication skills, particularly when it comes to structuring and expressing ideas independently.

More significantly, developments such as deepfakes, emotional influence and the difficulty AI has in interpreting real world contexts, raise important questions regarding trust and authenticity.

“WHILE AI HAS THE ABILITY TO STREAMLINE PROCESSES, REDUCE TIME SPENT ON TASKS AND SUPPORT DECISION-MAKING, IT LACKS THE CONTEXTUAL UNDERSTANDING AND PROFESSIONAL JUDGEMENT REQUIRED IN MORE COMPLEX OR AMBIGUOUS SITUATIONS”

While AI has the ability to streamline processes, reduce time spent on tasks and support decision-making, it lacks the contextual understanding and professional judgement required in more complex or ambiguous situations.

From a teaching perspective, this reinforces the idea that technology alone does not create value and that human oversight remains essential.

Professional scepticism and critical judgement

A key concern arising from the use of AI in auditing and accounting is its impact on the development of professional scepticism.

The concept of professional scepticism is central to the accounting profession, describing the state of maintaining a questioning mindset and critically evaluating information, rather than accepting it at face value.

While many students take steps to verify AI-generated outputs through manual checks and by cross-referencing, this is not consistent across all submissions.

There is a risk that students may accept AI outputs too readily, particularly as they are often presented in a confident and structured manner. This creates a false sense of accuracy, making it more difficult for students to identify errors or inconsistencies.

Some students only become aware of their reliance on AI when reflecting on an assignment following its completion. While many report that AI enhances their learning, there is also clear acknowledgement that it reduces their independent critical thinking.

This suggests that over-reliance on AI can develop gradually and without students fully recognising it. The concern is not that students are using AI, but that they may begin to depend on it in a way that limits their engagement with the material they are studying, and their ability to think independently.

This issue is particularly relevant in accounting, and especially in areas such as auditing and sustainability reporting, where judgement and interpretation are key. Sustainability reporting, for example, often involves qualitative information, which requires careful judgement-based evaluation and the application of professional scepticism.

If students are not encouraged to question and validate AI-generated outputs, there is a real risk these essential skills may not fully develop.

In an article published in December 2024 in the Harvard Business Review, authors Martin Reeves, Mihnea Moldoveanu and Adam Job of BCG Henderson Institute, the US consulting firm, argue that decisions are not merely exercises in data aggregation and algorithmic analysis, but involve selecting trustworthy data sources, applying judgement and considering the feasibility of outcomes.

This reinforces the importance of ensuring students do not rely solely on AI outputs, but instead develop the ability to question, interpret and critically evaluate information independently.

AI conflict in sustainable finance

Students have identified a potential conflict between sustainability and the use of AI. While they recognise that AI could support the analysis of large sustainability reports, they are also aware of the environmental impact associated with its use.

This introduces an additional challenge, as using AI responsibly requires both technical understanding and an awareness of its broader implications.

They highlight prompting as particularly important, saying poorly structured prompts often result in less accurate or overly broad outputs.

This suggests the effective use of AI is not as straightforward as it may appear, requiring a level of skill in prompting and refining outputs. From a teaching perspective, this raises the question: are students being supported in the development of these skills, or simply experimenting through trial and error?

There is also a concern that the routine use of generative AI for written work may lead to a loss of individuality.

This is important, as students need to develop their own voice and professional judgement. If AI is overused, there is a risk that student work could become standardised, reducing originality and limiting the development of communication skills.

Some students acknowledge using AI to fully complete some tasks—such as designing and generating posters, for example.

At the same time, they maintain that they add their own input elsewhere, but what does “elsewhere” mean?

This highlights a blurred boundary between the use of AI as a support tool and its use in the completion of substantial portions of academic work. From a teaching perspective, this raises questions about where this boundary should be drawn.

Capability replacement versus support

Students explicitly state that AI could produce more professional outputs than their own work, particularly in areas such as visual presentation.

This introduces an important consideration: students may begin to see AI not just as a support tool, but as a replacement for their own capabilities in certain areas. This could impact confidence in their own skills and reduce motivation to develop them further.

Further, AI systems lack foresight, moral reasoning or the ability to operate effectively in qualitative or uncertain environments.

While these systems perform well in structured, quantitative contexts, they cannot make ethical or professional decisions. This reinforces the need for human oversight and highlights the limitations of relying on AI in areas requiring judgement and ethical consideration.

Cost is another practical challenge. Many AI tools operate on subscription-based models, and not all students have equal access to advanced features. This raises questions regarding fairness and accessibility, particularly where AI becomes embedded in learning and assessment.

Overall, while AI clearly enhances efficiency, supports understanding and improves the presentation of work, it also introduces significant challenges. These include overreliance, reduced cognitive engagement, ethical concerns relating to bias and transparency and the potential erosion of professional scepticism.

These challenges must be considered carefully in the context of teaching practice to ensure AI is used in a way that supports learning, rather than replacing it.

Student interaction with AI is not uniform, but varies across engagement, understanding and awareness, reinforcing the need for clearer guidance and structured integration in teaching.

Implementing AI in accounting assignments

Students must develop a solid grounding in core accounting, financial and sustainability principles. While the use of AI should be embraced as a support tool, its integration into learning requires careful consideration. There are clear risks associated with over-reliance, including potential material misstatements, data privacy concerns, misinterpretation of outputs, weak cross-referencing and insufficient verification.

In addition, excessive dependence on AI can undermine critical thinking, reduce motivation and limit deeper engagement with the subject matter.

In their 2019 paper, Artificial intelligence in education: promise and implications for teaching and learning, authors Wayne Holmes, Maya Bialik and Charles Fadel argue educational institutions must ensure AI is embedded within strategy, rather than treated as an add-on.

In an educational setting, this means designing curricula and assessments that integrate AI in a way that supports learning objectives.

If students are not actively encouraged to question and validate AI-generated outputs, there is a danger these essential analytical and evaluative skills may not fully develop.

Transparency has always been central to education, and this remains unchanged. Proper citation and disclosure of AI use is critical. AI should enhance, not replace, independent thinking and final judgement.

Students should therefore be required to clearly explain their reasoning, calculations and conclusions in their own words to demonstrate a genuine understanding of the subject-matter at hand, rather than relying solely on AI-generated results.

Assignments should be structured in a way that ensures students engage meaningfully with the material rather than relying on AI to complete tasks.

Understanding, reasoning and evaluation

From my perspective, this requires a shift in assessment design, where the focus moves from output to understanding, reasoning and critical evaluation. This can be achieved in several ways:

1Case study-based assessment Case study-based assessments can be used, requiring students to analyse financial or sustainability reports and justify their conclusions independently, ensuring they engage with the underlying content.

2“Explain your workings” approach Similarly, the “explain your workings” approach, whereby marks are awarded for reasoning rather than just the final answer, can help students demonstrate a clear understanding of what they have prepared.

3Critical reflection on AI use AI can also be incorporated more deliberately into assessments. Tasks where students use AI and then critically reflect on the accuracy and limitations of the output can encourage a more inquiring approach.

4Data verification exercises In the same way, data verification exercises, in which students are required to identify and correct errors in AI-generated reports, can help develop professional scepticism.

5Scenario-based assignments Scenario-based assignments are also important, particularly in an accounting context, as they require students to apply professional judgement and consider ethical implications. Oral presentations can further support this by requiring students to defend their analysis and demonstrate genuine understanding of their work.

6Comparative analysis and peer review Comparative tasks, where students evaluate AI-generated responses against authoritative sources, can help reinforce the importance of validation and critical thinking. In addition, group projects incorporating peer review, can encourage accountability and discussion, reducing the likelihood of reliance on AI.

7Citation in research assignments Finally, research-based assignments should continue to emphasise proper citation and transparency, including the disclosure of AI use.

8Problem-based learning tasks Problem-based learning tasks that require interpretation, rather than straightforward calculation, are particularly valuable in ensuring students engage with the material at a deeper level.

From a People, Process, Technology (PPT) perspective, while AI capability is advancing rapidly, the corresponding development of people and processes does not always keep pace.

Accounting students are becoming increasingly confident in their use of AI tools, developing this technical capability at a fast pace.

This is not always matched by the development of critical thinking, professional scepticism and independent learning, however.

This imbalance presents a risk that AI may be used effectively from a technical standpoint without actually delivering meaningful learning outcomes. As a result, the focus in accounting education must shift towards ensuring people and processes evolve alongside the technology.

From an ethical perspective, this imbalance also gives rise to concerns regarding over-reliance on AI among accounting students, alongside reduced originality and the potential erosion of professional judgement.

Similarly, important legal considerations—particularly in relation to data privacy, transparency and the appropriate use of AI-generated content—require clear guidance and oversight within educational settings.

As the use of AI among students continues to rise, the lecturer’s governance role becomes increasingly important. Lecturers have a responsibility to ensure AI is used appropriately, transparently and in a way that supports learning.

This includes designing assessments that encourage critical engagement, requiring students to validate and justify their work and ensuring the use of AI is disclosed where appropriate.

The abiding value of human oversight

The role of the accountant is evolving. As AI takes on more aspects of routine tasks and basic computations, the profession is shifting towards a more strategic advisory and consultancy role.

In their 2016 MIT Sloan Review article, co-authors Thomas Davenport and Julia Kirby write that cognitive computing solutions are designed to enhance human decision-making rather than replace it, enabling professionals to focus on more complex, judgement-based tasks.

Accountants will still need to interpret information, communicate insights effectively and apply their expertise in real-world contexts.

Human oversight will remain essential, particularly in navigating ethical considerations, applying professional scepticism and ensuring compliance with regulations.

Although some entry-level roles may diminish, new opportunities are likely to emerge that are more intellectually engaging and value driven.

In her 2014 book The Good Jobs Strategy, for example, MIT Sloan School of Management Professor Zeynep Ton, outlines how “the smartest companies invest in employees to lower costs and boost profits”.

It is no longer sufficient to ensure students understand accounting and finance; they must now also be able to communicate this knowledge, apply it in practice and use AI effectively without becoming dependent on it.

This requires students to develop a clear understanding of the nuances and limitations of AI, while also strengthening their ability to exercise professional judgement and critical thinking in increasingly complex, technology-enabled environments.

Linda Mc Weeney is a Senior Lecturer in Accounting and Finance at Technological University Dublin