AI Auditing

A necessary tool to strengthen responsible AI and provide assurance

Authors

  • Dirk Brand Stellenbosch University
  • McEllory Hoffmann Praelexis AI
  • Johan Van der Merwe Praelexis AI
  • Aloisio Dourado Neto TCU

Abstract

This article reflects the results of a research project to develop an AI Audit Framework for Supreme Audit Institutions, focusing on public sector usage. It addresses the need for transparency, fairness, accountability, and alignment with ethical and legal requirements. The article draws on international standards and best practices to offer a methodology for auditing AI algorithms across their entire lifecycle, including risk categorization, data governance, and bias assessment. It also details how generative AI poses new challenges that require specialized guidelines. Recommendations highlight interdisciplinary collaboration and continuous skill development to ensure responsible AI governance.

Author Biographies

Dirk Brand, Stellenbosch University

Independent legal consultant, special counsel at Swart Law and an Extraordinary Senior Lecturer at the School of Public Leadership, Stellenbosch University. He holds the following degrees: BComm, LLB, LLM (European Union Law) and LLD (Constitutional Law). He was the leader of the LegalAIzers team who won the First St.Gallen Grand Challenge in Switzerland on the application of the EU AI Act (July 2023). He is also a guest lecturer at the Hochschule Kehl, Germany, the Law Faculty, University of Verona, Italy, the SSML Pisa, Italy and West University Timisoara, Romania. He serves as a Policy Fellow at the Centre for Artificial Intelligence and Digital Policy (CAIDP) in Washington D.C. Much of his current work focuses on AI ethics, AI policy and AI assessment.

McEllory Hoffmann, Praelexis AI

CEO and co-founder of Praelexis, a machine learning company that empowers its clients to become data-driven decision makers. He is also an extraordinary senior lecturer in Computer Science at Stellenbosch University. He has a BEngSci (Cum laude) in Computer Engineering and a PhD in Applied Mathematics, from Stellenbosch University. As a qualified computer engineer, mathematician and data scientist, McElory has discovered the different ways in which AI can solve a business’s unique challenges. He is fascinated by how technology can make the world a better place and has devoted his life to helping others utilise AI to create a sustainable future. 

Johan Van der Merwe, Praelexis AI

Data Science Strategist at Praelexis. He has the following degrees: BA, B.Th. and M. Th. in Systematic Theology, as well as an MBA and a PhD in Ethics, all from Stellenbosch University. His current focus areas include business model design and business strategy, as well as data strategy and agile methodology. He is also a Research Fellow at Stellenbosch University

Aloisio Dourado Neto, TCU

Bachelor's degree in electrical engineering from UFBa, a master's in applied computing focused on data analysis, and a Ph.D. in Computer Science with a focus on AI and Computer Vision from UnB. A public servant at TCU since 2010, he has been working with data analysis for auditing since 2014. Currently, he is the Deputy Chief Auditor for Data Analysis and Innovation at the Secretariat of Strategic Information for Control, where he is responsible for the Court's Data and Analytics Governance.

Downloads

Published

2026-07-17

Issue

Seminário Internacional O Futuro da Auditoria Pública

Section

Eixo 3 - Abordagens inovadoras na auditoria do setor público: casos e lições aprendidas