INTHEBLACK April 2026 - Magazine - Page 42
F E AT U R E
tools and processes, with industry bodies supporting
training and providing networks of people or experts
to share their knowledge,” said D’Cruz.
Consequently, there is an opportunity to explore
the interplay between ethics as it relates to climate
change and climate change accounting at an
undergraduate level.
“Universities are trying to catch up really quickly
in the ethics space and it is evolving,” said Stokes.
“I think organisations need the skill set now; otherwise,
there will be issues around assurance. Ethics will always
have inherent societal evolution risks.”
then why are you charging? For us, it is a really
challenging time, but it is exciting. AI helps
to improve audit in general, but do not trust the
AI to do the entire check. Balance that technology
component with your responsibility as a professional.”
Using AI for routine accounting tasks like
safeguarding confidentiality and signing off activities
carries considerable risks, in large part due to the
technology’s lack of transparency, which heightens
the need for ethical oversight.
“The question of ethics comes down to the AI
models,” said D’Cruz. “They can often be a black box
and we do not have control of what the model initially
spits out, although we can heavily review and vet this
output before the user sees it.”
A large language model (LLM) is a type of AI, such
as ChatGPT. LLMs draw on banks of information to
create content. What is material when it comes to this
technology is how the information an LLM produces
is used within a business.
“Our boundaries with AI are very much related
to our values, so we have to make sure it is not
hallucinating and vet and train all our models
to make sure they output the right data,” D’Cruz
added. “We have hundreds of AI researchers, scientists,
machine learning engineers and data scientists globally
researching and reviewing the model outputs.”
HUMANS REMAIN ESSENTIAL
As it stands, AI is not a flawless technology — humans
are still required to provide substantial oversight.
Yatsuzuka believes this will continue. “In my view of
the future, the humans still need to be there,” he said.
“In a hundred years’ time, CPAs might be doing
different things but there will still be a system working
for the general public.” ●
AI AND TRUST
The accounting profession is still sorting through
the ethics of AI’s application, the boundaries of which
will continue to evolve as the technology matures.
At the heart of the issue is trust — how much
confidence should be placed in AI outputs and
how much oversight it needs.
“You always need that element of professional
judgement when reviewing work,” said Maggie Lo CPA,
financial operations manager at Employment Hero.
Yatsuzuka agrees. “At the end of the day it is our
advice, not the AI’s advice. So we have to ensure we
put all our resources into really getting on top of it.
If you cannot add anything additional to the output,
42 INTHEBLACK APRIL 2026 SPECIAL EDITION
READ
an article about navigating the ethics of generative AI
in business
EXPLORE
CPA Australia’s range of ethics short courses