There are no statutory limits on using AI to create or check valuations meaning professional judgement must always be employed with this practice.
Sector practitioners face no restrictions in using artificial intelligence (AI) to generate or assess the market valuation of an asset but an over-reliance may be considered a breach of professional obligations, Smarter SMSF chief executive Aaron Dunn has stated.
Speaking at the ASF Audits Technical Summit 2026 in Adelaide late last week he said AI was likely to have a future rile in creating and verifying valuations and noted the only limits on it was how practitioners should be using its output.
“There is nothing that would prohibit the use of an AI tool helping us in some way, shape or form to come up with the basis for providing information to come up with a value,” Dunn noted.
“What I would then add is that AI is the tool that will help you to build it, it is not the source of what you ultimately want to create.”
“The value in what you get out of it is how you hit it not in terms of just what it spits out.
“So if I wanted to try and justify [a valuation] I could tell [AI tool] Claude to do that until the cows come home and it will build a fantastic case to justify that figure, but that is Coronica in the modern day,” he added, in reference to case heard before the Administrative Appeal Tribunal in which a trustee incorrectly valued assets to avoid the inhouse asset limit.
“You are using a tool, and in case of Coronica [v Commissioner of Taxation [2024] AATA] the tool was the trustee’s own knowledge and experience, but you are trying to create an argument for something that you know is fundamentally wrong.
“You have to still be able to know where to hit it, to be able to justify and build a file that would support the external evidence that might come with it.
“This is the third party reasoning where you dig up stuff the trustee or you might not be able to see and make sure you get to a point you can document something that would ultimately become explainable to a third party.”
Dunn indicated using AI to assemble, collate, cross-check, and monitor information was safe use of the technology but going to the opposite extreme and allowing it to produce values and invent comparables had to be avoided.
“The models will confidently produce a sale of an asset for a particular value that actually never happened,” he explained.
“You will have come across that if you have used AI as it currently stands, and because it looks like evidence it would naturally sit in the file but again it’s just Coronica with better formatting.”
Jason is a senior journalist with Benchmark Media and writes for the selfmanagedsuper website and magazine and smstrusteenews website. He has covered financial services since 1999 and has written about life insurance, superannuation, investment management and financial advice, in both a freelance and in-house capacity, and also did a brief stint as a media manager for an industry association. He has been the editor of trade titles Money Management and Financial Standard, and was most recently senior journalist with life insurance title Riskinfo.
PO Box 96
Lindfield NSW 2070
Contact
Advertise
selfmanagedsuper is the definitive publication covering Australia’s SMSF sector. It uniquely offers online content tailored separately for SMSF professionals and individual trustees participating in the fastest growing and largest sector of the superannuation industry. As such, it is a must read for those wanting to stay informed about the latest news, regulatory developments, technical strategies, investments, compliance, legal and administration issues concerning SMSFs.
Copyright © SMS Magazine 2026
ABN 80 159 769 034
Benchmark Media
WordPress website development by DMC Web.
Sign up for the latest news updates

source