‘My super is in cash’: Ex-RBA economist’s AI warning
A former RBA economist has issued a brutal new warning to Australians, revealing the big reason his superannuation is “entirely in cash”.
The AI boom is driving investment returns for millions of Australians, but one expert isn’t buying the hype.
Martin Eftimoski, an AI researcher at the University of New South Wales and former economist at the RBA, is worried the more than $150 billion that big tech spent on Artificial Intelligence this year won’t yield the promised results.
And he warns Australians might be overexposed, as funds like AustralianSuper invest heavily in tech giants.
Mr Eftimoski is so concerned about a potential crash that earlier this year, he took a drastic step with his own superannuation.
“My super is entirely in cash and has been for a few months,” the lecturer told news.com.au.
“I just feel like a real return of two to three per cent over the next few years is fine for me. “That sounds a lot better to me than the chance I get wiped out by 30 to 40 per cent have to start from scratch.”
Mr Eftimoski was quick to add that he wasn’t advocating other Australians should follow his example. Historically shares have yielded greater returns than cash and he acknowledged he would “probably” lose money as a result of his decision.
The S&P 500 index has soared to new all-time highs in recent months, driven up by mega-cap tech companies like Apple and Microsoft, which are leading the AI revolution.
But amid optimism that AI will turbocharge productivity and drive earnings growth even higher, there’s also concern the market is overvalued.
MORE: How AI is helping Aussies find work
“You’re seeing massive amounts of capital expenditure in data centres, and that’s creating the impression that all this expenditure will generate returns.
“The thesis is if you keep building a bigger and bigger model, it just becomes God, I suppose, but that doesn’t seem to be true.”
Last month, OpenAI founder Sam Altman said he believed investors as a whole were “overexcited” about AI.
“When bubbles happen, smart people get overexcited about a kernel of truth,” Mr Altman said in an apparent reference to the dot-com bubble, when investor euphoria about internet-based companies led to the Nasdaq losing almost 80 per cent of its value in the early 2000s.
Investor jitters were heightened when MIT released a study which found 95 per cent of AI projects had showed no returns on investment for companies.
AI bulls argued the research had relied on narrow metrics, however, and may have overlooked the impact of employees using AI on their own terms - and beyond the oversight of their managers - in a phenomenon dubbed “shadow AI”. The technology’s potential could take years, if not decades, to fully unlock, with hiccups and naysayers as much a part of the process as the adoption of railways, computers or the internet.
A bear in a bull market
For Mr Eftimoski, a combination of the AI boom, political instability and institutional decay in the US has left him much more comfortable holding cash.
He is deeply sceptical about the race for businesses to adopt the current darling of the AI space, the large language model (LLM).
MORE: The highest paying jobs in Australia without a degree
“I don’t think the majority of businesses understand how to employ AI, and I don’t think the majority of managers understand what it is,” he said.
“The business excitement around LLMs is far greater than the actual usefulness. Look at it from a personal perspective: have you ever been lied to by ChatGPT?”
LLMs can give the false impression of intelligence, objectivity and confidence - all while fabricating answers in a bid to please the user or continue a pattern that seems statistically likely, even if it’s factually wrong. In short, they can flat-out lie, a problem referred to in the industry as “hallucinations”.
Mr Eftimoski pointed to an example from 2022, when Air Canada’s chatbot was caught promising a discount that didn’t exist.
The chatbot assured a customer that he could book a full-fare flight for his grandmother’s funeral and then apply for a bereavement fare after the fact.
A tribunal later ordered Air Canada to pay the customer damages, after the airline had tried to argue the Chabot was a “separate legal entity that is responsible for its own actions”.
“The technology itself is designed to meet your perceptions. That’s why it’s so good at therapy,” Mr Eftimoski explained.
“It’s designed to meet you where you’re at, and that is why it’s very good at sneaking its way into projects, into your life. Business leaders are enamoured with it because it tells them what they want to hear.
“The reality is that people don’t always want to hear the truth, and the truth is what you need.”
