Focus: AI is rewriting the rules of accounting education

Rapid advances in artificial intelligence are transforming accounting skills and assessments, as well as the way universities prepare future professionals. This calls for new approaches to teaching, learning and student support in an AI-enabled environment.

The 2025 SAICA and University of Cape Town Accounting Education Symposium brought together academics and industry partners to examine how artificial intelligence is reshaping the profession and what this means for professional accounting education.

AI has already changed how accountants work and make decisions. The real challenge now is helping education keep pace without losing the principles that underpin trust in the profession. Future accountants will need strong technical fluency with AI as well as the critical human-centred judgement that remains at the core of good professional practice.

The symposium opened with a welcome from Professor Suki Goodman, Dean of Commerce at UCT, who spoke about the strength of the university’s accounting legacy and the pressures created by rapid technological change. South African firms are already using machine learning tools to detect anomalies, interpret unstructured data and support audit procedures.

For Goodman, this means universities must rethink curriculum design and assessment, and the capabilities expected of graduates. The goal is not only to keep pace with technology but to ensure that core values such as scepticism, integrity, analytical judgement, ethics and continuous learning form the backbone of good practice. She reminded attendees that the workplace has already shifted and that universities need to meet it where it is rather than where it used to be.

SAICA’s Executive for Capabilities and Endorsements, Riyaan Mabutha, highlighted that the profession is entering an exciting era of transformation. Disruptive technologies, the rise of sustainability reporting, and growing expectations for ethical leadership are redefining what it means to be a chartered accountant. Mabutha emphasised SAICA’s cradle-to-career philosophy as a key strength and noted that collaboration between universities, regulators, and employers will unlock new opportunities for relevance and impact. He encouraged the profession to embrace AI as an enabler rather than a threat, explaining that technology will enhance accountants’ roles rather than replace them. To meet future demands, academia and industry must work together to equip graduates with the ability to leverage AI tools, interpret outputs, and apply sound judgement in complex scenarios − skills that will position them as trusted advisors in a rapidly evolving world.

SAICA Chief Executive Patricia Stock described the decade ahead as a defining period shaped by globalisation, rapid technological acceleration, shifting public expectations and a more competitive global talent market for accounting and finance skills.

Artificial intelligence, she noted, is no longer just an automation tool. It generates content, asks questions, summarises complex information into workable insights, and interacts with users in increasingly sophisticated ways. Stock argued that the central issue for the profession is how to preserve trust while adopting technology at scale. Graduates need to understand how to use AI, but they also need to lead with humanity and ethical clarity. They need to know when to override automated results, when to escalate concerns and when to apply insight that AI cannot provide. She called for curriculum rationalisation to avoid overwhelming students with technical volume and stressed the importance of mentorship and professional guidance.

Stuart Pedley-Smith Founder and lead consultant at Synapse Learning Solutions opened the ‘AI and the Accountant of the Future’ session by explaining how large language models are reshaping accounting competencies. Routine tasks will increasingly be automated, shifting the accountant’s role towards strategic interpretation and advisory support.

Future professionals will need strong analytical thinking and the confidence to evaluate and challenge machine-generated information. Pedley-Smith also highlighted the opportunities of AI-enabled education, including adaptive learning, simulations and automated feedback, while warning about risks such as misinformation and digital inequality.

Hackeem Dante Hafkey and Clint McLean of Audit AI showed how students are already integrating AI deeply into their learning. Students use AI for explanations, summaries, study planning and practice questions. Many combine several tools to build surprisingly advanced workflows. However, this has created a digital divide between those who can afford premium models and those who rely on free tools. They urged institutions to create clear AI policies, ensure equitable access, upskill academics and build consistent guidance so students understand when and how AI can be used responsibly.

Richard Mellon and Nic Pullen from Groundflr concluded the first session with insights into AI-driven grading. Their work demonstrates that AI can process scripts faster and apply rubrics with consistent logic. In some cases, AI applied reasoning more precisely than human markers. However, AI grading introduces challenges such as managing inference, preventing hallucinations and ensuring strong quality control. They too urged institutions to create clear policies around using AI in assessments, ensure equitable access, upskill lecturers and establish strong governance frameworks to manage risks such as bias, misuse and uneven learning outcomes.

Turning to the practical uses of AI in teaching, Jana Lamprecht from the University of the Free State presented a Bot-a-thon activity where students design functional bots to solve real problems. The exercise builds digital acumen and encourages experimentation. Students created tools ranging from email summarisers to sales-data organisers.

Pieter Pienaar and Dr Charisa de Klerk from the University of Pretoria demonstrated AI simulations designed to build communication and emotional intelligence by mimicking the demands of real client interactions where students must integrate audit and tax knowledge under pressure. Student feedback suggested the simulations deepened understanding and boosted confidence more effectively than written tasks.

A team from UCT introduced SmartMarker (Chris Guattari-Stafford, Antony Rogers, Ilse Lubbe, Magon Gajewski, Jacqui Dean (UCT)), an AI-powered assessment platform that integrates digitisation, rubric-based scoring and human review. They expanded on the technical and ethical demands of automated marking, including the difficulty of interpreting handwritten responses, ensuring consistency across large cohorts and preventing algorithmic errors from going unchecked. Strong governance structures, detailed audit trails and agent-based workflows allow SmartMarker to maintain accuracy and accountability but still leaves final judgement in the hands of the academics.

The final session turned to curriculum transformation. The presenters (Chair Kim Krause; panellists: Hamman Schoonwinkel (SUN), Sylvia Banda (UL), Cameron Modisane (Unisa), Nico Strydom (UJ), Jesse Manickum (NWU)) agreed that accounting education needs to move away from content-heavy learning and towards competency-based models that focus on integrated problem solving.

AI should be woven through the curriculum rather than treated as a separate topic. Speakers called for flexible structures, adaptive assessments, strong industry partnerships, better lecturer training, stronger digital skills, clear AI policies and more authentic learning experiences that reflect real professional work.

The consensus was that AI is enhancing the work done by accountants, but core values, such as ethics and trust as well as technical competence still define it. Education must adapt to these emerging tools and trends, but it must also strengthen prospective chartered accountants digital capabilities rather than diluting them.

Author
Monique Verduyn