The AI Ethics Guide Accountants Were Told to Expect Is Here
This article will count 0.25 units (15 minutes) of unverifiable CPD. Remember to log these units under your membership profile.
You run a client's numbers through an AI tool. The output looks right. You move on. But did you actually understand how that result was reached? Could you explain it to a regulator? According to new global guidance published in July 2026, that question is now central to your professional obligations.
The International Ethics Standards Board for Accountants (IESBA) has released a new staff publication titled Emerging Technologies: A Characteristics-Based Approach to Ethical Considerations for Professional Accountants. It applies to all professional accountants, whether you are in public practice or working inside a business, and it covers AI, machine learning, generative AI, robotic process automation, distributed ledger technologies, and quantum computing.
The document is a direct message to practising accountants: the technology you use does not reduce your ethical responsibility. It increases the care you need to take.
Why this publication matters now
South Africa's accounting profession has been rapidly adopting AI-driven tools for bookkeeping, tax preparation, financial reporting, and client work. The IESBA's 2023 technology-related revisions to the Code came into effect in December 2024. This new publication builds directly on those revisions, offering practical guidance on how the characteristics of emerging technologies create specific ethical risks under the Code.
This publication is also the first in a series. The IESBA has confirmed that more technology-specific guidance will follow, with the next publication expected to focus specifically on artificial intelligence.
The seven technology characteristics you need to understand
Rather than focusing on individual tools, the IESBA takes a characteristics-based approach, designed to remain relevant as technology continues to evolve. Seven characteristics are identified that can create or amplify threats to your compliance with the Code's fundamental principles.
Opacity means many AI systems operate as a "black box." The system produces an output you cannot fully explain. If you rely on a result you cannot explain to a client, regulator, or colleague, you may be breaching your obligations around integrity, objectivity, and professional competence.
Non-determinism means the same inputs can produce different outputs at different times. This makes it harder to reproduce audit trails or give consistent advice. You cannot simply trust a result you cannot verify.
Data dependence means AI outputs are only as reliable as the data behind them. Poor, biased, or incomplete data creates biased results, and if your advice is based on those results, you carry the risk.
Perpetual adaptivity means AI systems can change their behavior over time based on new data and interactions, without you necessarily being told. A tool that worked correctly six months ago may now produce different results.
Autonomy means that when a system makes decisions with little human involvement, accountability gaps open up. The IESBA is clear: regardless of how automated the tool is, the professional judgment and the accountability remain yours.
Scalability and speed together mean errors and bias can spread across large volumes of work faster than any human can catch them.
The speed of AI output also creates pressure to accept results without proper scrutiny, which the IESBA explicitly names as a threat to objectivity.
The fraud risk you may not be thinking about
Emerging technologies have made fraud significantly easier and more convincing. The IESBA publication specifically highlights these risks:
Criminal actors can now use AI to create deepfakes, which are realistic fake videos, images, and audio that convincingly imitate real people or events.
AI can generate synthetic data and other fabricated content that is difficult to distinguish from genuine source material.
Documents, representations, and supporting information that you rely on in client work can now be manipulated in ways that were not previously possible.
The reliability of outputs and evidence you receive can no longer be taken for granted.
The IESBA's message is clear: professional skepticism now needs to extend explicitly to the authenticity of the information and evidence you receive, not just the conclusions you draw from it.
What this means for business accountants in practice
The guidance applies throughout the entire technology lifecycle, from when you choose a tool, to when you implement it, use it, update it, and eventually replace it. It is not a once-off assessment.
For accountants in practice, this means you need to be able to explain the outputs of every tool you use in client work. You also need to monitor tools over time. A tool that passed your initial assessment may behave differently after a vendor update, and the Code requires you to reassess when circumstances change.
For accountants in commerce, the same principles apply. AI-powered dashboards, forecasting tools, and reporting systems all carry the same risks. Sufficient human oversight is not a nice-to-have. It is an ethical requirement.
A practical starting point is to review your technology stack and ask three questions about each tool: Can I explain how it reaches its outputs? Do I know whether it has changed since I last assessed it? Am I maintaining enough human oversight to catch errors before they affect my work or my clients?
The accountants who thrive will be those who combine the efficiency of AI with the judgment and ethical discipline that no tool can replace.
Read our related publications: AI in the accounting profession; 7 rules every accountant must know before using AI